Who's Who in Text Analytics

Text analytics has become increasingly important for many use cases but the market consists of numerous vendors many of which are using immature technologies. Our Who's Who will help IT leaders find a vendor to suit their particular needs.

• Powerful trends like social media and e-discovery have driven IT leaders in business analytics (BA) and enterprise content management (ECM) to look at text analytics.
• Most text analytics offerings support a single use case well but vendors often claim more general capabilities.
• A small number of factors keep text analytics from being the simple, powerful solution that many users want. They include immature technology, difficulty meshing with a larger ecosystem, the diversity of use cases and the fact that text analytics is not a single market.

• If text analytics serves a strategic purpose in your organization, create a centralized program to enable an efficient and coordinated approach to implementation.
• If text analytics are only for specific tactical purposes look for niche, "best-of-breed" vendors to support different use cases rather than trying to find products that can solve all problems or attempting to force unsuitable products do things they aren't meant to do.
• For external social media analysis you should look mostly toward the providers of software as a service (SaaS) solutions.
• Evaluate a range of vendors – at least half a dozen vendors support each of the major use cases, such as e-discovery, voice of the customer (VoC) and social media analysis.
• Allow additional time and financial resources and demand a comprehensive proof of concept from vendors that do not offer references from companies of a similar size to your own.
• Organizations should plan to acquire the necessary skills for the effective deployment of text analytics and assign people trained in the use of these technologies. Your business intelligence (BI) competency center should be the ideal department to embrace these skills and allow a coordinated approach to the deployment of text analytics combining BI for even larger benefits such as hybrid analytics.

Powerful trends in social media, e-discovery, customer services (call center transcriptions of voice calls, customer complaint emails and instant messaging) and customer-centric business strategies are driving IT leaders in BA and ECM to consider text analytics as a powerful business tool. Text analytics can churn through large volumes of documents to derive business insight, understand customer behavior, automate processes and organize information. However, the text analytics market consists of a large number of vendors, many of them small, which are using immature technologies and products that may not be suitable for some uses. A couple of megavendors like IBM and SAS do offer comprehensive text analytics which are far from simple to use but offer technical that are pretty mature. To help IT leaders reduce the number of vendors on their shopping lists, Gartner has created this Who's Who of text analytics players differentiated by market segment and what their products can and can't do.

This research follows the market segment definition in "Market Profile: Text Analytics" which is based on how the IT professional is likely to encounter vendors and products.

Vendors are differentiated by the following classifications:

• Megavendors are defined by their revenue size, distribution channel strength, installed base, and focus on business transactions and information management.
• ECM and search platform vendors either use text analytics as the core technology base of their search functionality or have added it as a component of their ECM offering.
• Pure-play platform vendors provide all or some of the three basic technical architecture stage technologies through full application platforms, development environments, SaaS, or plug-in components through APIs.
• Application specialists provide an application with text analytics as a feature and not as the defining functionality (an e-discovery product that includes text analytics as one of many required modules, for example).

The vendor selection in this research was based around companies that our clients or the creators of this research have personal knowledge of. This means they have either been included in client inquiries, have been present at events or have shown up on the radar of a Gartner analyst in the recent past. Gartner asked each of the featured vendors to respond to a short RFI about their product's out-of-the-box text analytics technology. The product has to be a stand-alone solution which can be either delivered on-premises, through the cloud or via SaaS. It should have the ability to read text through a variety of mechanism. The output then had to be reused in a many different ways. Companies that did not return their surveys were excluded at this time as there is no way to confirm the correctness and completeness of the information given here (see Note 2). This research is not intended to be a comparative evaluation of the products discussed here, nor is it intended to be an exhaustive list. As such, this research does not cover all available text analytic vendors or products. This research is based on information from verified RFI response and coverage in previously published Gartner research.

Five main factors keep text analytics from becoming the powerful solution to many needs that users and vendors wish for:

1. Immature technologies and products
2. Difficulty fitting into a larger ecosystem
3. Diverse use cases
4. Short supply of sophisticated skills both in IT organization and the end user community
5. Text analytics is not a single market

Immaturity: Most text analytics offerings support a single use case well but vendors often claim more general for their platforms. Many vendors that describe their offerings as generic really focus primarily on e-discovery. Even the completeness and configuration of these offerings varies widely. Many vendors do not offer out-of-the-box products but a variety of tools from which customers have to build their own solutions. Some of the products on the market have commodity functions that will not differentiate an enterprise from its competitors. Finally, enterprises cannot easily apply the most advanced products to their intended tasks. Text analytics requires new types of skills to leverage the more advanced and most valuable features of these products, such as combining insight from social media content with customer orders to optimize campaign results and capture new opportunities as they emerge.

Ecosystem: IT leaders will usually want to use text analytics to help with BA or ECM tasks. In both cases, three challenges are preventing text analytics from thriving within these ecosystems:

1. Most text analytics offerings do not easily integrate into existing BA platforms. Vendors provide APIs which require a lot of work to integrate the tools in the way people want to use them.
2. Text analytics technologies have a history in the government intelligence and security markets that the vendors are not allowed to talk about because of non-disclosure agreements. Police, internal audit organizations, intelligence and law enforcement agencies have been using the text analytic technologies that most ECM buyers view as cutting edge today for a long time.
3. New sources of data – which are mostly unstructured and include emails, customer call center interactions, customer transcripts, documents, machine-generated data and social media data – come into the enterprise with such volume and velocity that they create new challenges for data integration and information management infrastructure. Additional storage and processing, which go way beyond the existing data warehouse architecture paradigm, will have to be considered. The fact that many of these providers are moving deeper into the cloud further exacerbates the situation.

These factors have prevented text analytics from gaining traction even where the technologies have a track record.

Use cases: The different uses cases that text analytic offerings support give their deployments divergent characteristics. For example, many vendors have VoC, social analytics or monitoring as specialties since they have domain-specific ontologies, but these technologies tend to be the newest and least mature – more tools than products. Some vendors can point to large-scale deployments but they almost all categorize content.

These three factors increase the challenge of evaluating text analytics vendors and offerings. The disparities make comparisons difficult and what IT leaders learn about text analytics from one vendor does not necessarily carry over to others. Until the market and its technologies mature, IT leaders should look at different vendors to support different use cases rather than trying to find one product to solve all problems while exercising appropriate caution over tool proliferation.

Short supply of skills: Because the supply of people with appropriate skill sets always lags behind technology adoption, acquiring skills or hiring people with relevant skill sets is a challenge. The proliferation of text analytics further exacerbates the demand/supply gap for both end users and IT support.

Text analytics is not a single market: As pointed out in "Market Profile: Text Analytics" text analytics is not a single market because it has multiple buying centers. Text analytics offers functions or features that are suitable for many industries as well as horizontal functions. IT professionals are facing the challenge of taking common text analytics and offering services that can be applied to multiple business-specific applications.

www.ibm.com

Product: IBM has a number of products that fit into the text analytics framework. Content analytics combine with enterprise search to provide additional insights with views and dashboard into ECM repositories. IBM's products and solutions use natural-language processing (NLP) technologies for improved understanding and classification of content (call center notes, social media, open-ended surveys and emails, for example).

IBM SPSS Modeler Premium (a data-mining workbench) includes text analytics functionality to help users combine structured and unstructured data (operational data stores, as well as files, emails, call center notes, blogs and RSS feeds in different languages) for predictive models.

IBM SPSS Text Analytics for Surveys is a stand-alone desktop product used in conjunction with IBM SPSS Statistics and IBM SPSS Data Collection which is based on NLP technologies specifically designed for survey text. It's used to transform unstructured survey data into quantitative data and for sentiment analysis.

IBM Content Analytics with Enterprise Search is a converged content analytics and enterprise search product based on NLP which can extract facts, concepts and relationships from unstructured content. It targets use cases in areas of public safety, crime intelligence, insurance operations, patient care and enterprise fraud.

IBM Content Classification provides the ability to automate content classification in analytics, search, document imaging and document repository from a document corpus.

IBM eDiscovery Analyzer provides capabilities to analyze and prioritize electronically stored information (ESI) to gain case insight using key text analytics embedded from IBM Content Analytics and enhanced to identify concepts and phrases, understand facts and communication threads and locate key pieces of evidence and witnesses.

The products offer:

• Data mining algorithms used for automatic classification and clustering
• Modeling algorithms to combine and visualize insights
• The ability to recognize, analyze and display patterns, trends, deviations and connections across a wide variety of sources by using a different algorithms
• Integration with a wide IBM portfolio, including Cognos BI and Infosphere, as well as external sources through support of the Unstructured Information Management Architecture and Lucene index
• Extensive APIs for input and output of results across a wide selection of sources and target systems

IBM also announced the successful acquisition of Vivisimo on 29 May 2012, a discovery and navigation application that helps organizations to find insights across structured and content data.

Focus: Content Analytics focuses on improving search results. The SPSS product focuses on predictive decision making and adding unstructured data to structured data models. IBM focuses on:

• Insurance
• Healthcare
• Public safety
• Banking
• Telco
• Retail
• Manufacturing
• Consumer packaged goods

Geography: Global presence.

Sales Model: Direct, partnerships and OEM relationships.

Table 1. IBM Capabilities

IBM
Information Extraction	Yes
Extraction Methods	NLP-based: Concept Named entities Fact/event/sentiment extraction
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	Yes (RDF and XML)
Information Visualization
Sector Adaptation	Yes
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Combination of multiple products for different outcomes Content Analytics is combined with Enterprise Search to drive better classification Watson and SPSS Modeler add additional functionality allowing IBM to cover almost any business use case
BI = business intelligence; NLP = natural language processing; REST = representational state transfer; RDF = resource description framework; SaaS = software as a service; SPSS = statistical package for the social sciences; XML = extensible markup language

Source: Gartner (September 2012)

www.sap.com

Product: SAP BusinessObjects Data Services uses NLP to extract, categorize and summarize information from free-form text sources such as email, documents and notes as well as data feeds from social platforms, such as Twitter.

(SAP also resells Social Media Analytics by NetBase, a SaaS social media insight and analytics platform that combines NLP analytics with a content aggregation service and a reporting capability. It contains one year of social media from blogs, tweets, newsfeeds and other Web content.)

The product:

• Automatically identifies and tags named entities such as people, companies, places, weapons, addresses and dates
• Identifies events (such as mergers and travel) concepts and sentiments as well as relationships between entities
• Allows taxonomy-based, document-level classification of information
• Integrates with SAP BusinessObjects Enterprise. SAP BusinessObjects Data Integrator software directly accesses text analysis sources and jobs. In addition, data quality software can post-process extracted entities such as performing data normalization or clustering mentions.
• Structured entities or facts from text data processing can be queried or reported on from Crystal Reports, BusinessObjects Web Intelligence software, NetWeaver technology platform, dashboards and other BA clients.

Focus: Integrating and embedding social media insight and analysis into SAP analytics products. Use cases include:

• Analysis of customer interactions in call centers and online customer chat sessions
• Evaluation of brand awareness and marketing campaigns from social media

Geography: Global presence.

Sales Model: Direct and partner sales.

Table 2. SAP Capabilities

SAP
Information Extraction	Yes
Extraction Methods	Core extraction (predefined entities in 11 languages) Custom dictionary and fact (relationship) extraction (rule-based)
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (31)
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	No
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization
Sector Adaptation	No
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	No
Special Connector	Yes
Web Services	No
Subscription-Based SaaS or Cloud Option	No
Open Source	Tightly integrated with Hana to enable high performance and real-time text analytic delivery Text analytics integrated in Data Services as part of data integration and data quality
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

www.sas.com

Product: The Text Analytics suite includes SAS Text Miner, SAS Enterprise Content Categorization, SAS Ontology Management and SAS Sentiment Analysis. SAS offers single-user versions, Text Miner for Desktop and SAS Content Categorization, which can operate on the desktop and can include any of the associated add-on modules.

The products can:

• Discover topics and patterns within entire document collections by mining unstructured data sources
• Organize, classify and extract concepts, entities and facts from documents to improve understanding and relevance of corporate data
• Generate metadata to create or enhance existing search, content management systems and document repositories with add-ons that offer prebuilt taxonomies, Web crawling, search and indexing, document deduplication and text summarization
• Provide consistent, systematic linkage across text repositories using semantic relationships
• Automatically locate and analyze the sentiment in electronic text in real time from websites, internal files and reports, surveys, forms, emails and communication centers to spot trends and identify customer priorities

Text Analytics supports Java, C, Python C#, .NET, PERL and Web Service APIs.

Focus: Integrating text-based information with complex event-processing systems and with structured data for use in predictive modeling, analytics and data management. The suite is also utilized in a number of industry-specific solutions with predefined reports, data models and domain expertise.

Use cases include:

• Quality control diagnosis from patient records
• VoC for effective customer communication solutions and intelligence
• Insight from social media to improve ROI for advertising
• Predicting analysis on trends and future problems

Geography: Global presence.

Sales Model: Direct sales.

Table 3. SAS Capabilities

SAS
Information Extraction	Yes
Extraction Methods	Supervised, semi-supervised and unsupervised techniques (NLP and statistical methods) Prebuilt industry taxonomies, SAS Industry Taxonomy Rules Wikipedia-deduced categories and sub-categories Customer-defined linguistic rules Fact and entity extraction rules and categories and subcategories deduced from Wikipedia using the DBpedia community project Text-mining derived topics (unsupervised classification) and complex Boolean relationships
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	No
Special Connector	Yes
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Integrated SAS Business Analytics Framework, analysis of textual data in conjunction with Data Integration, Business Intelligence and Predictive Analytics Combining linguistic parsing with statistical models to achieve better classification with disambiguation, and entity recognition Hybrid models combining both linguistic and statistical methods as well as an ensemble model can each improve classification and accuracy over one single method Active learning with machine-derived Boolean relationships refined by end user High-performance processing for addressing big data
BI = business intelligence; NLP = natural language processing; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.ai-one.com

Product: NathanApp; NathanNode and NathanCore (formerly Topic-Mapper) is an API-based platform technology that enables text analytics in any language. It's technology mimics the way in which the brain detects pattern to find high-order co-occurrences and identify latent relationships between data elements across systems. In addition to machine learning technology, the company also builds reusable user interfaces into the mechanics and results of the machine learning process. It can also work as a stand-alone tool or in conjunction with external ontologies, natural language processing, databases and analytic tools. ai-Cloud and ai-Search provide intelligent semantic search, news feed analysis and query enhancement for content providers, publishers and websites. The ai-Browser plug-in enables concept search from any source or target text for analysts and researchers in enterprise and government. These products integrate with all BA vendors that support XML feeds.

Focus:

• Genomic sequencing and forensic solutions
• Pattern recognition of profiles, behavior and semantics
• Generation topic maps from social sites such as Twitter or talk radio

Geography: Presence in Europe and North America.

Sales Model: Direct sales, partnerships.

Table 4. ai-one Capabilities

ai-one
Information Extraction	Yes
Extraction Methods	XML compliant dictionaries Taxonomies Ontologies Extraction rules
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	OWL, RDF, Microformat
Information Visualization	No
Sector Adaptation
Out-of-the-Box	No (NathanCore)
Required Third Party	Yes (NathanApp and NathanNode)
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	NathanCore No NathanShell Yes
Claimed Differentiators	Machine learning based on neurophysiology – trademarked "biologically inspired intelligence" Learns by association
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.autonomy.com

Note: Autonomy was bought by HP in 2011. Autonomy founder Mike Lynch left HP in May 2012.

Product: Autonomy IDOL 10 extracts meaning from all forms of information, including audio, video, social media, email and Web content, as well as structured data such as customer transaction logs and machine-based sensor data. Autonomy's Stratify, Introspect, Investigator and Early Case Assessment products address every stage of the e-discovery life cycle, spanning the complete Electronic Discovery Reference Model. Records Manager addresses information classification, retention and disposition. Autonomy sells software licenses as well as appliances for e-discovery, archiving and enterprise search.

Focus: Autonomy provides software infrastructure to manage information, including all forms of text data. Autonomy also provides solutions for:

• E-discovery
• Compliance
• Legal
• Content management

Geography: Global presence.

Sales Model: Direct sales and through HP's reseller and distribution channels.

Table 5. Autonomy Capabilities

Autonomy
Information Extraction	Yes
Extraction Methods	NLP Conceptual Contextual
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	No
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Contextual understanding Understand concepts Language independent Learns continuously Part of broader IDOL platform containing 500 functions (including clustering, classification and entity extraction), 400 connectors and 1,000 file formats Packaged content analytics apps (including e-discovery, customer intelligence, social media and sentiment analytics)
BI = business intelligence; NLP = natural language processing; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.opentext.com

Product: OpenText's Content Analytics:

• Extracts information from a wide array of content repositories and enterprise- or user-generated content
• Uses thesauruses, taxonomies, ontologies, syntactical patterns and machine learning to establish facts and relationships across multiple domains
• Uses this information to create additional navigation possibilities, monitor applications, automatically classify documents and evaluate content
• Supports REST and XML remote procedure call input and output of information so that it can link it to a wide variety of systems

Focus: Classifying and structuring content for horizontal use cases, particularly records management and augmented search, with no industry-specific uses. However, OpenText does offer Content Analytics as part of prepackaged products available for:

• News and media
• Business
• Insurance and financial services
• Education
• Life sciences and healthcare

Geography: Global presence.

Sales Model: OpenText sells Content Analytics as an on-premises product as well as a SaaS offering.

Table 6. OpenText Capabilities

OpenText
Information Extraction	Yes
Extraction Methods	Controlled vocabularies Computational linguistic Rules
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes
Workflow	No
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	Yes
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Text analytics are packaged into middleware Packaged solution Semantic search Automatic classification for records managers
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.pingar.com

Product: Pingar API consists of 18 text analytics components and provides in three areas:

• Rapid discovery (content search and management)
• Entity extraction (information extraction)
• Content analysis (analysis of text context)

Pingar API is SharePoint-friendly and integrates with forensic solutions.

Focus: Government, legal and manufacturing.

Use cases include:

• Uniformly structuring state legislation and making it searchable by topic
• Adding metadata and structure to unstructured documents in a SharePoint environment

Packaged solutions are also available for:

• Compliance
• ECM
• E-discovery
• Healthcare
• Scanning

Geography: Asia/Pacific focus, limited presence in Europe and North America.

Sales Model: Direct sales, OEM and partnership. Pingar API is available on-premises or via on-demand cloud services. Pingar levies fixed monthly charges based on number of calls to API per day and documents per call. A free version (limited to 1,000 calls per day) is available for developers, students, small Web projects, non-profits and proofs of concept.

Table 7. Pingar Capabilities

Pingar
Information Extraction	Yes
Extraction Methods	Entity extraction Postal addresses Taxonomical relations
Sentiment Analysis or Opinion Mining	No
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	No
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	No
Concept Relevance	No
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	No
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	No
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Invoice analysis (detecting important parts of a tax invoice for further automated processing) Automated sanitization and redaction of input documents (for removing sensitive content)
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.smartlogic.com

Product: Smartlogic's Semaphore content intelligence platform consists of several modules that can be implemented independently or integrate with customer applications. Smartlogic also offers implementation services including an information scientist to work with the customer's domain experts to accelerate model building. The Semaphore platform employs a range of semantic software technologies in combination with domain-specific vocabularies (taxonomies and ontologies) to add a layer of meaning and context to unstructured information.

Features include:

• Taxonomy and ontology management
• Text analysis and extraction
• Metadata management
• Automatic classification
• Semantic reasoning and processing
• Context-aware navigation

Integrations: Semaphore has resilient, scalable, XML-based Web service interfaces that allow the model and classifications to integrate with content management systems, enterprise search engines, business process management systems and workflow applications where semantic markup is needed.

Focus: Smartlogic is targeting "information heavy" industries such as government, healthcare, advisory, research, finance, media and high-tech manufacturing. Use cases include the automatic distribution of information based on content, such as automatic routing of alerts to the relevant disease officer based on reports received from healthcare and medical centers. Some blue chip companies use this capability to improve their search engine results and deliver more relevant content.

Geography: Europe, North America.

Sales Model: Direct sales.

Table 8. Smartlogic Capabilities

Smartlogic
Information Extraction	Yes
Extraction Methods	Ontology- and taxonomy-based extraction Custom (and highly sophisticated) rule-based extraction Named entity extraction using a combination of algorithms and specialized dictionaries
Sentiment Analysis or Opinion Mining	No
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (32)
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	Yes
Special Connector	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Ability to determine and then provide metadata – the values of which are derived from any type of text – to any consuming application that requires it The provision of semantic services to provide a contextual user experience over information for any type of user experience (desktop, tablet or phone for example). The ability to comprehensively process textual content for the purposes of compliance and control
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

www.zemanta.com

Product: Zemanta is a SaaS-only Semantic Web extension that provides add-ons for a number of blogging platforms. Zemanta analyses text as it is being typed and suggests additional content that can be linked as well as relevant images that match the content. Zemanta uses text analytics across a wide range of domains to match the terms it encounters against a database of recognized terms and then make recommendations. It suggests tags and categories, and uses its own index to provide additional content for linking in real time. Domain-specific ontologies and Zemanta's own ontology provide the backbone for the linking.

Zemanta has made its API available to a number of commercial installations that use its technology.

Focus: Zemanta does not offer a domain-specific product or service but instead provides its API calls to customers. Its widest adoption is as part of a number of popular blogging platforms, and it is used by a number of commercial news services and their blogger networks to improve on-content linking.

Geography: Focus in Europe with a limited presence in North America.

Sales Model: The product is SaaS-only and commercial clients pay by the number of API calls they make during a specific time.

Table 9. Zementa Capabilities

Zementa
Information Extraction	Yes
Extraction Methods	Out-of-the-box ontology Entity database
Sentiment Analysis or Opinion Mining	No
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	No
Workflow	No
Document Summarization	No
Organizing and Structuring
Categorization	No
Clustering	No
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	No
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Cross-domain analysis
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.alchemyapi.com

Product: AlchemyAPI is a text mining platform featuring semantic analysis capabilities in the natural language processing field. AlchemyAPI is a SaaS text and content analysis API and there is an appliance version. AlchemyInsight uses the text analytics technology for big data visualization.

AlchemyAPI:

• Analyzes WordPress blogs and suggest new categorization tags for posts
• Uses text mining and an NLP engine for semantic-powered search engine optimization
• Processes any RSS or ATOM feed and enhances partial RSS feeds into full content feeds
• Annotate feeds with named entities, topics and other metadata
• Performs NLP and text analysis on Linux or Unix systems
• Supports Java, C/C++, C#, Perl, PHP, Python and Ruby
• Supports ai-one solutions

Focus: AlchemyAPI focuses on:

• Social media monitoring
• Advertisement targeting
• Real-time search and discovery
• Tracking of influencers and sentiment within the media
• Content aggregation and recommendation
• Stock trading decisions
• Business and government intelligence systems
• Recommendation engines
• Search engine optimization

Geography: North American focus.

Sales Model: Direct sales. AlchemyAPI has five tiers of service from completely free to enterprise-scalable solutions with service-level agreements. Each tier has a specific limit on the number of calls per month.

Table 10. AlchemyAPI Capabilities

AlchemyAPI
Information Extraction	Yes
Extraction Methods	Custom dictionaries/lexicons Custom ontology/taxonomy
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	No
Workflow	No
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Comprehensive functionalities
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

www.attensity.com

Product: Attensity Command Center delivers real-time social analytics in a view that shows at a glance what customers are saying right now about a company, products and brands. Attensity Analyze extracts insights from customer conversations from a variety of social, online and internal sources for understanding VoC. Attensity Respond helps enterprise social media teams and customer service teams listen and respond to customers communicating via social media such as tweets, Facebook posts, forums, blogs and communities, as well as over email, on website, chats, letters and faxes. Users can organize conversations into queues based on the content of the message then route them to the right individual to respond.

Attensity products:

• Categorize sentiment based on patented technology called "Exhaustive Extraction TM" which combines multiple NLP approaches
• Analyze information about geography, demographics, influencers, issues and root-cause analysisv
• Offer over 100 out-of-the-box reports
• Integrate with Amdocs, Microsoft, Nice, salesforce.com, SAP and Siebel (Oracle)

Focus: Attensity text analytics applications serve marketing and customer service users in:

• Telecommunications
• Financial services and insurance
• High-tech industries
• Consumer electronic and white goods
• Hospitality and travel

Use cases include:

• Analyzing customer feedback to proactively reduce customer churn
• Analyzing customer service records, emails, survey responses and online community forums to find the root cause of customer complaints
• Amassing customer feedback, inquiries and content which contains data about customer sentiment, service issues and "cries for help" that are hard to analyze and react to
• Large-scale customer engagement response across social media and more traditional channels

Geography: Focus on North America, the U.K. and the DACH regions (Germany, Austria and Switzerland).

Sales Model: Direct sales and through strategic partners like Capgemini and Cognizant.

Table 11. Attensity Capabilities

Attensity
Information Extraction	Yes
Extraction Methods	Holds nine patents around text analytics including: Entity extraction Classification Language identification Geo-location Exhaustive extraction Machine learning Event extraction Anaphora resolution Topic disambiguation
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (32)
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	No
Clustering	No
Concept Relevance	No
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Analysis of social media feeds across 150 million sites in real time Deep NLP Deep root-cause analysis Designed for business users
BI = business intelligence; NLP = natural language processing; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.clarabridge.com

Product: The Clarabridge suite provides text and sentiment analysis from a number of different sources including local files such as emails, surveys, contact center notes and public social sites such as Facebook and cloud services such as salesforce.com. Clarabridge uses a combination of rule-based categorization, theme detection, clustering and machine learning to achieve its results. Its sentiment analysis allow for a granular understanding of topic, degree, context and domain sentiment even when the languages are mixed.

Clarabridge has out-of-the-box integration with SAP BusinesObjects, IBM Cognos and MicroStrategy and other platforms can be added.

Focus: Customer experience management and VoC for organizations that focus on brand monitoring or have a large external audience with needs, sentiments and discussions which need to be understood. Clarabridge offers around 20 different industry templates including those for retail, hospitality and online banking.

Geography: Europe and North America.

Sales Model: Perpetual license for an on-premises installation or a SaaS model around the volume and the number of users.

Table 12. Clarabridge Capabilities

Clarabridge
Information Extraction	Yes
Extraction Methods	APIs Database extraction Reports Completely open analysis scheme Collaborate module Microsoft Office suite
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (9)
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Root-cause analysis Comparative analysis reporting Real-time alerts Satisfaction scoring Theme detection
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.contentanalyst.com

Product: Content Analyst Company has been in the market since 2004 and markets its Content Analyst Analytical Technology (CAAT) platform primarily as an OEM product. It uses a patented technology called Latent Semantic Indexing, which is language independent and can identify any Unicode language, where content is mathematically analyzed, compared and concepts extracted. It does not use taxonomies, word lists or dictionaries but rather relies on a representation of extracted concepts learned from the material itself and a small number of reference documents. Proximity is then assumed to represent similarity thereby creating the possibility of categorization. The CAAT platform provides functions and features supporting dynamic clustering, concept-based categorization/search, near-duplicate document detection, email analytics, instant context and summarization and social media monitoring.

All of these functions are offered as tools to be used by an exclusive OEM channel or integration partner that can then use the APIs to build additional I/O mechanisms, user interfaces, or integrate the technology into its own platform.

Partner companies include: kCura, iCONECT, iPro, Mindseye, Driven, TCDi, Altep, IPStreet, Decooda, Agilex, AnyDoc and Benchmark Intelligence.

The platform offers:

• Mathematical algorithms for concept-based organization
• Modeling algorithms to combine and visualize insights
• Conceptual search including real-time context
• Dynamic clustering
• Extensive APIs used by OEM partners to integrate CAAT into a variety of their own products

Focus: Content Analyst Company focuses around categorization of information in documents, Emails and other large bodies of data for:

• Information governance
• E-discovery
• Forensics
• IP patents
• Intelligence communities
• Federal public sector
• Digital marketing
• Social media monitoring

Geography: Global presence.

Sales Model: Indirect sales to OEMs and system integrators.

Table 13. Content Analyst Company Capabilities

Content Analyst Company
Information Extraction	Yes
Extraction Methods	Mathematical model-based Entities Metadata
Sentiment Analysis or Opinion Mining	No
Big Data Capabilities	No
Multiple Language Support (Number of Languages)	Yes, any Unicode language is supported
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	No
Information Visualization	No
Sector Adaptation	Yes
Out-of-the-Box	CAAT is a toolset, functionality depends on OEM vendor
Required Third Party	Yes
Development Tools or Standard Interface for Application Integration Real Time Capabilities Predictive Analytics Integration	Yes Yes Yes Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Cross-lingual search Language Identification Instant context concept-based keyword expansion Difference highlighting

Source: Gartner (September 2012)

www.collectiveintellect.com

Product: Collective Intellect, which was acquired by Oracle on 5 June 2012, offers a SaaS solution that combines internal enterprise text analytics with real-time monitoring of external conversations. The main products are CI Brand Intelligence, which tracks conversations around a brand and CI Brand Action, which integrates with CRM, business process management and business analytics applications and recommends direct actions.

Collective Intellect:

• Uses real-time scrapers to feed several pipelines with data
• Uses analytics to score, categorize, detect the sentiment and deduplicate the information scraped to increase accuracy
• Uses latent semantic analysis and NLP to understand the conversation and assign a sentiment to it
• Can include structured and unstructured text from within an organization to enrich and complement its findings and present them in various dashboards
• Provides a REST API as a data feed to make its findings available to other programs

Focus: SaaS product is used to track conversations and identify influencers in social media channels. The vendor is expanding the product to include triggers for actions, multichannel engagement and VoC solutions.

Collective Intellect focuses on:

• Consumer packaged goods
• Financial services
• Media and entertainment

Geography: Focus in North America.

Sales Model: Offers several pricing models including fixed fees, price per message and basic, premium or professional packages.

Table 14. Collective Intellect Capabilities

Collective Intellect
Information Extraction	Yes
Extraction Methods	Categorization Enrichment Snippetization Dimensions Scoring
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	No
Multiple Language Support (Number of Languages)	No
Workflow	No
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Speed and accuracy of processing Consumer intentions analysis Share of voice analysis White space analysis
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.expertsystem.net

Product: Expert System products offer text analytics, automatic document categorization, content tagging, search, ontology and taxonomy management and automated self-help solutions. Its products include Cogito Search Explore Engine, Cogito Intelligence Platform, Cogito Categorizer, Cogito Discover, Cogito Semantic Tagger, Cogito Answers and Cogito Studio which are based on a single core technology.

Features include:

• Out-of-the-box extraction of named entities and type-based elements such as dates, measures, addresses and complex entities
• Ability to identify an entity and the associated features (such as being able to predict a user's age)
• Ability to discover the relationships connecting two or more entities (between a driver and his or her car, for example)
• Sentiment analysis of individual elements
• Products integrate with IBM Cognos, SAP BusinessObjects and iQube

Focus:

• Knowledge management
• Customer care and intelligence
• Automated Q&A for self-help solutions in the oil and gas industries
• National intelligence agencies
• Automatic categorization and tagging of editorial content
• Media in publishing
• Media in telecommunications

Geography: Focus in Europe and a limited presence in North America.

Sales Model: Direct sales and through value-added resellers, OEMs, integration partners and consulting firms.

Table 15. Expert System Capabilities

Expert System
Information Extraction	Yes
Extraction Methods	Named entities and associated features Type-based elements Concepts Identification of complex relationships
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support	No
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	Yes
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	Yes
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	The only text-mining platform based on deep parsing and a multiple language semantic network Includes multiple language support for sentiment and categorization Has the ability to remove the inherent ambiguity in language and understand words in context
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.knime.com

Product: Knime Desktop is an open-source business intelligence platform specializing in text mining.

Features include:

• Enterprise support and maintenance service for this product as well as Knime Professional, Knime Team Space and Knime Server
• Over 1,000 native and embedded nodes covering various data loading, transformation, analysis, and visual exploration models as well as "all you can think of" text-mining functions and features
• An open API that can be integrated with other Eclipse projects such as BI and reporting tools and data tools platforms to provide more functions

To get the most out of Knime products organizations will need the services of analytical professionals with in-depth knowledge of text mining and analytics as well as IT skills.

Focus: Customers that want to differentiate themselves through text analytics and by combining text analytics with other type of analytics, such as network analytics.

Knime customers operate in:

• Life sciences
• Telecommunications
• Banking
• Police departments
• Retailers of fast-moving consumer goods

Geography: Worldwide.

Sales Model: Knime Desktop is available under GPL license. Offers one-year licenses for support and maintenance from direct sales, distribution partners and consulting partners.

Table 16. Knime Capabilities

Knime
Information Extraction	Yes
Extraction Methods	Enrichment Transformation Preprocessing
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language	Yes
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	No
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	Yes
Claimed Differentiators	Comprehensive and powerful tools to support the development of different kinds of text analytics applications Integration with networking and data mining
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.lexalytics.com

Product: Lexalytics provides the Salience text processing engine through a set of callable APIs which are typically used to integrate into another platform. Sister company Semantria provides Salience via SaaS. Entity Management Toolkit and Sentiment Toolkit enable users to configure a new statistical model for extracting specific types of entities and combining them with sentiment phrases.

Features include:

• A set of libraries with API wrappers for all common programming languages including C, C#, Java, Python and PHP
• An array of text analytics functions including summarization, concept/theme extraction, sentiment analysis, named-entity detection, categorization and conceptual understanding

Focus: Lexalytics:

• Acts as an OEM to a variety of technology companies
• Powers many of the Web's leading social media platforms and feeds
• Supports use cases such as sentiment analysis, content classification and enhancement of search results

Geography: North America.

Sales Model: Direct sales.

Table 17. Lexalytics Capabilities

Lexalytics
Information Extraction	No
Extraction Methods	Entity extraction Categorization (import taxonomy mapped onto a concept matrix) User-defined rules
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	No
Workflow	No
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	No
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real-Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	Yes
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Extreme configurability
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

www.alias-i.com/lingpipe

Product: LingPipe API is a set of Java development tools for processing text using computational linguistics.

Features include:

• Multilingual, multidomain and multigenre models
• Training with new data for new tasks
• N-best output algorithms with statistical confidence estimates to optimize search
• Online training (learn-a-little, tag-a-little)
• Thread-safe models and decoders for concurrent-read, exclusive-write synchronization and character encoding-sensitive input/output
• Categorization of Twitter search results and suggested spelling corrections of queries
• Integration with StreamBase CEP

Focus: Legal and professional firms.

Use cases include:

• Query spell checker (using the West Law search engine and various Thomson news services)
• Named entity detection for classified deployments
• Sentiment analysis and language identification
• Tag cloud summarization technology for websites to identify interesting projects for funders

Geography: North America.

Sales Model: Online and direct sales.

LingPipe packages include:

• A free download with use restrictions
• A per-server developer license
• A per-server startup license for small companies
• An enterprise server license with full support services
• Packaged tools based on task and training parameters are also available

Table 18. LingPipe/Alias-I Capabilities

LingPipe/Alias-I
Information Extraction	Yes
Extraction Methods	Named entity detection Predicate argument extraction Basal noun phrase extraction Database linking or normalization
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	No
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	No
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real-Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	No
Special Connector	Yes
Web Services	No
Subscription-Based SaaS or Cloud Option	No
Open Source	Yes
Claimed Differentiators	Mature code base with broad use in commerce and research Not a shrink wrap solution Ideal toolkit for highly customized applications Excellent documentation and teaching resources
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

www.provalisresearch.com

Product: Provalis Research offers two text analytics products: QDA Miner provides computer assistance for manually coding, annotating, retrieving and analyzing small and large collections of documents and images; WordStat is a text analytics module for QDA Miner. It combines content analysis methods by using a dictionary approach and text mining methods.

Features: QDA Miner integrates statistics and visualization. It includes geo-tagging and time-tagging tools that provide:

• On-screen coding and annotations of texts and images
• Computer assistance for coding with more than seven text search tools
• Integrated statistical and visualization tools
• Integrates with Provalis Research's SimStat for statistical analysis and development of more complex applications

WordStat features:

• Integrated exploratory text mining and visualization tools to automatically extract themes and identify trends and patterns
• Automatically categorize unstructured information by using existing or creating new hierarchical content analysis dictionaries composed of words, word patterns, phrases and proximity rules
• Computer assistance for dictionary building
• The ability to relates unstructured text with structured information

Focus:

• Government and nongovernmental organizations
• Universities
• Marketing research
• Security
• Insurance
• Manufacturing

Geography: North America.

Sales Model: Direct and online sales (licenses are available for commercial, government and academic use).

Table 19. Provalis Research Capabilities

Provalis Research
Information Extraction	Yes
Extraction Methods	Dictionary based (taxonomies or ontology) User-defined dictionaries composed of words Word patterns and phrases Proximity rules for measurement of concepts
Sentiment Analysis or Opinion Mining	No
Big Data Capabilities	No
Multiple Language Support (Number of Languages)	No
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes (optional SDK)
Real-Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	No
Open Source	No
Claimed Differentiators	Integration of WordStat with QDA Miner allows content analysis, the text-mining features of WordStat, manual coding and the annotation of QDA Miner to be combined in many ways Integrated analysis and visualization including multidimensional scaling plots, dendrograms, heat maps, word clouds and correspondence plots Advanced geotagging and time tagging tools allow the creation of dynamic maps and interactive timelines
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; SDK = software development kit; XML = extensible markup language

Source: Gartner (September 2012)

www.rapid-i.com

Product: Rapid-I is one of the only open-source text analytics vendors (enterprise support is available). Its main products are RapidMiner and RapidAnalytics. RapidMiner focuses on text analytics and predictive analytics. RapidAnalytics is a cloud analytics service for visualizing and creating interactive reports.

Rapid Miner's features include:

• Statistical and analytical actions such as sentiment analysis
• Categorization using naïve Bayes methods, decision trees and neural networks expanded with numerous clustering methods and machine learning
• Direct import of information and linking to many different files types as well as connectors to databases and SAP enterprise resource planning

In addition, a community of users have developed a large number of extensions to the products. Both products run in a Java runtime environment.

Focus: Large and small organizations with no focus on particular industries or domains.

Geography: Global.

Sales Model: Partner and direct sales. All products are available via subscription, perpetual or OEM licenses and as cloud or on-premises editions.

Table 20. Rapid-I Capabilities

Rapid-I
Information Extraction	Yes
Extraction Methods	Dictionaries Regular expressions XPath
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support	Yes
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real-Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	Yes
Claimed Differentiators	N/A
BI = business intelligence; N/A = not available; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.saplo.com

Product: Saplo offers automated text analytics functions in form of its Tags, Match, Sentiment and Context APIs that can extract text content, including names and entities, based on the structure of text without the need for a reference list:

• Tags automatically defines the meaning of words and extract entities such as companies, people, locations and topics from texts.
• Match is a service that understands the meaning of text and the relationship between phrases with similar meanings allowing it to present related and interesting content to users for further exploration.
• Context is a service that predicts an outcome based on one or several texts.
• Sentiment uses semantic technologies to extract and evaluate feelings and emotions expressed by writers.

Saplo products use semantic models that interpret text in a similar way to humans with a system that automatically learns from user feedback. It is language independent.

Focus: Supports large and small information providers and media companies. Use cases include:

• Contextual advertisement
• Text recommendations
• Spam filtering
• Brand monitoring
• Search engine optimization
• Media and website optimization
• Risk reduction

Geography: Europe.

Sales Model: Direct sales and in partnership with content management vendors.

Table 21. Saplo Capabilities

Saplo
Information Extraction	Yes
Extraction Methods	Entities Topics
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	No
Multiple Language Support	Yes
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real-Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	Yes
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Text prediction services Saplo Match and Saplo Context refine hidden values in text data Feedback system in Saplo Tag and Saplo Sentiment Language independent No ontologies needed for new languages
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.temis.com

Product: The Luxid Content Enrichment Platform is typically used through Web services as part of larger workflows, often in connection with a content management system. Off-the-shelf integrations include those for SharePoint 2010, EMC Documentum, Alfresco and Marklogic.

The platform includes the following key modules:

• Luxid Annotation Factory uses NLP technologies to enrich unstructured content with additional metadata, content linking and categorization in order to create structured information.
• Luxid Content Enrichment Studio is a suite of four productivity tools that enables the customization and custom development of Skill Cartridges (see below) focusing on new domains and the evaluation and tracking of extraction quality.
• Luxid Information Analytics makes enriched content visible, usable and comparable via search, navigation and dashboard widgets.
• Luxid Skill Cartridges extend Annotation Factory with extractors, focusing on topics such as biology, medicine and chemistry, competitive intelligence and opinion mining, legal concepts and International Press Telecommunications Council categorization.

Focus: Temis takes a generic platform approach.

Key customer groups include:

• Publishing
• Life sciences
• Homeland security

Use cases include:

• Linking discoveries in scientific data
• Creating links across multiple content repositories to allow wider relationship building among content objects
• Discovering insights in large bodies of data such as scientific literature, customer feedback or news articles

Geography: Europe and North America.

Sales Model: On-premises and deployments supported by perpetual or annual licenses.

Table 22. Temis Capabilities

Temis
Information Extraction	Yes
Extraction Methods	Linguistic rules (morphology and syntax) Semantic rules (sentence patterns) Dictionaries, thesauri and taxonomies Statistics Machine learning Business rules
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (20)
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real-Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Full range of NLP techniques enables easy entry points and highly sophisticated implementations Capable of extracting relationships between entities in addition to topics and terms Comprehensive tools suite enables use in virtually any use case or domain Broad multilingual capability and distributed architecture to support scalable and multinode deployments Off-the-shelf integrations available for Microsoft SharePoint 2010, Alfresco, EMC Documentum, Drupal, Nuxeo and Marklogic
BI = business intelligence; NLP = natural language processing; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.discovertext.com

Product: DiscoverText is a cloud-based toolkit which analyzes text documents, such as open-ended answers on surveys, emails, Excel spreadsheets and social media to help businesses make better marketing, service, operational and product decisions. DiscoverText requires partial manual coding to train the active-learning engine.

Features include:

• Custom machine classification through human coding and machine learning
• Latent Dirichlet algorithms
• Naïve Bayesian techniques
• Cloud explorer for interactive word classification
• Unlimited access to Twitter data via the GNIP PowerTrack (optional)

Focus: Texifter provides text analysis for:

• Employee engagement surveys
• Marketing and branding
• VoC
• Legal e-discovery
• Federal public comments
• Academic research

Geography: North America.

Sales Model: Online and direct sales. DiscoverText comes in a Professional Edition, with all functions and 100GB of data storage; an Enterprise Edition with 5GB data storage; and a Community Edition which is free but is limited to three basic functions.

Table 23. Texifter Capabilities

Texifter
Information Extraction	Yes
Extraction Methods	Custom machine classification through human coding and machine learning Latent Dirichlet Algorithm
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	No
Multiple Language Support (Number of Languages)	Yes (5)
Workflow	No
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	No
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Coder adjudication Coder inter-reliability measurement tools Custom machine classification using proprietary technology Cloud-based SaaS solution Rooted in academic research Ten years of NSF-funded research Not limited to social media Valuable to customer and employee survey research teams Better data quality through custom classification and coder adjudication
BI = business intelligence; NSF = National Science Foundation; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.cicdata.com

Product: CIC currently has three products:

• IWOMdiscover is used for buzz tracking on all of China's major social media platforms including microblogs, social network services, news feeds, bulletin board systems, blogs, video sharing and Q&A sites.
• IWOMexplorer provides marketers and researchers with market intelligence for brand promotion, social media marketing, product management, customer service and consumer research.
• IWOMcooperator provides real-time analysis and engagement on modern social media, including the identification of influencers, multiaccount management and marketing management services.

An implementation service is required for on-premises installation. CIC also provides consultants for social media analysis.

Focus: Social media marketing, social media monitoring, analytics and marketing management.

Geography: Primarily in the People's Republic of China but also Asia/Pacific region.

Sales Model: Subscription by selection of topics, brands and entities; customized service on special topics, entities and brands are available.

Table 24. CIC Capabilities

CIC
Information Extraction	Yes
Extraction Methods	Proprietary semantic processes
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (3)
Workflow	Yes (with IWOMcooperator on social CRM)
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Large volume data coverage on social media sites in China Proprietary Chinese text mining technology Knowledge of data insights and social business gives constant improvement on insight delivery and expansion in social business support Big data capabilities for all historical data archive
BI = business intelligence; CRM = customer relationship management; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

Product: eMudhra offerings include:

• Veda Recruiter which aggregates candidate details from emails, websites and file systems, and extracts key information to match open jobs.
• Veda Social Analytics Platform provides real-time sentiment analysis from social media such as Facebook and Twitter on specific topics and concepts.
• Veda Patent Search provides search results for patents which are ranked and sorted based on relevance.
• Veda Business Workflow enables users to access contextual content seamlessly from within a business application without having to move between applications.

These offerings can be integrated into standard BA platforms.

Focus: Veda solutions target the recruitment business and research companies. Veda also automates the filing of tax returns, based on information extracted from hard-copy tax return to reduce manual data entry.

Geography: Asia, North America.

Sales Model: Direct sales.

Table 25. eMudhra Capabilities

eMudhra
Information Extraction	Yes
Extraction Methods	Entity recognition using visual segregation, NLP, dictionary and taxonomy Rule-based pattern recognition Lexical analysis Word stemming Phrase identification Text extraction from the most popular text document formats
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	No
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	Yes
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Entity extraction based on visual segregation does not depend solely on language processing techniques and relies more heavily on visual representation of content The representation of words in a different format from other words increases the importance of these words and these play a vital role in entity extraction Visual segregation entity extraction does not depend on similar template formats Platform differentiates itself from the machine learning approach as there is no need for the system to learn from a huge data corpus Supports unstructured and semi-structured text formats
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.nice.com

Products: Nice Systems has three main text analytics products that extract information from chat, email and social channels:

• Enterprise (for customer interactions)
• Security (for security organizations)
• Actimize (for financial risk and compliance)

In addition, Interaction Analytics translates speech into text for further analysis. The product allows for context, sentiment and topic analysis. It uses NLP and statistical models to present the results to perform root-cause analysis. Fizzback ties customer feedback to predefined categories by using NLP techniques.

Focus: Nice Systems products target:

• Enterprises that want to enhance customer experience through operational efficiency and better analysis of interactions
• Financial institutions that want to enforce compliance
• Government and private agencies dedicated to public safety

Geography: Global presence.

Sales Model: Fizzback is available only via SaaS. Nice offers Interaction Analytics in both SaaS and on-premises models.

Table 26. Nice Systems Capabilities

Nice Systems
Information Extraction	Yes
Extraction Methods	Extraction rules Dictionaries Statistical methods Semantic inference
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	No
Multiple Language Support (Number of Languages)	No
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	No
Real Time Capabilities	Yes
Predictive Analytics Integration	No
BI Platforms Integration
XML Feeds	No
REST API	No
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Real-time analysis for text-based conversations and SMS-based surveys Hot topics detection Real-time dashboard on current topics Identification of new emerging topics Topic trend overtime Interaction part analysis Automatic ML-based interaction segmentation Event extraction and reason analysis Web user session summarization
BI = business intelligence; ML= maximum-likehood; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.radian6.com

Product: salesforce.com's Radian6 platform provides real-time access to online social conversations including blogs, online forums, news, social networks, and insights into vendors such as Clarabridge, Klout, OpenAmplify, Peek Analytics, Solariat, Reuters OpenCalais and PeekYou.

Radian6 components focus on the following aspects of social analytics:

• Listening (Radian6 platform)
• Measuring (Summary Dashboard, Social Insights)
• Engaging (Engagement Console, Social Hub, Radian6 for the service cloud)
• Learning (Analysis Dashboard)

Features include:

• Extracting of information from social media and social insight providers
• Analysis of context, entity, topics and sentiment
• Rules and workflow
• Integration through a REST API

Focus: Enterprises that want to enhance customer relationship management and better handle public relationship through analysis of social conversations.

Use cases include:

• Brand awareness
• Customer services
• Public relations

Geography: North America, Europe, Latin America and Australia.

Sales Model: Subscription-based SaaS and reseller partnerships.

Table 27. salesforce.com Capabilities

salesforce.com
Information Extraction	Yes
Extraction Methods	Keyword Boolean Named entity Topic Key Noun Phrase Sentiment
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	No
Multiple Language Support (Number of Languages)	Yes (18)
Workflow	Yes
Document Summarization	No
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	Yes
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Integration with CRM Ability to action insight directly from applications via workflow Platform flexibility – Ability to add new partners with relative ease and then utilize their algorithms on the vast amounts of social data monitored by Radian6
BI = business intelligence; CRM = customer relationship management; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.sysomos.com

Product: Sysomos's monitoring software tracks and analyzes conversations in the social media arena and helps users to act upon them. Its main product, the Media Analysis Platform, monitors the main social media platforms via a dictionary enriched by its own algorithms. Heartbeat is a real-time monitoring tool used for tracking terms, brands or topics across all included channels.

Sysomos Media Analysis Platform:

• Has been optimized for the unstructured, informal and abbreviated language of social media like Twitter streams
• Uses statistical machine learning
• Can work in multiple languages
• Visualizes conversations and sentiment
• Integrates with text-based sources such as Twitter, Facebook, blogs, forums and newsfeeds
• Has APIs to push visualizations and analysis into customer intranets or Web pages

Focus: Sysomos focuses narrowly on the analysis of social media content. Sentiment analysis allows marketers and others to monitor these channels and watch for brands and influence leaders. Sysomos is not industry specific.

Geography: Focus in North America and Europe.

Sales Model: Sysomos is available only as a SaaS platform.

Table 28. Sysomos Capabilities

Sysomos
Information Extraction	Yes
Extraction Methods	Keyword lists Machine learning Dictionary assisted
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (5)
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	No
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	No
REST API	Yes
Special Connector	No
Web Services	No
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Designed for social media (informal, short text) Works with terabytes of data and billions of conversations Ease of use: Does not overwhelm the analyst with use of intelligent visuals
BI = business intelligence; REST = representational state transfer; SaaS = software as a service; XML = extensible markup language

Source: Gartner (September 2012)

www.verint.com

Product: Verint's main text analytics product Impact 360 monitors, captures and analyzes customer interactions and related sentiment across text-based communications channels, including survey responses, email, Web chat, agent notes, social media, salesforce.com and other cloud information sources. It serves as part of the company's VoC solution which captures feedback across multiple channels.

Impact 360's text analytics functionality integrates with Verint's enterprise feedback management (EFM) and the Verint speech analytics engine allows it to analyze voice as well, for example, in call centers or customer service departments. Impact 360 analyses content and presents the results so that users can act on them.

Impact 360 Text Analytics:

• Has more APIs available for additional integration to other information sources, including salesforce.com, Radian6, Lithium and Buzzmetrics
• Integrates with IBM/Cognos, SAP/Business Objects and Oracle Business Intelligence Enterprise Edition

Focus: Domain agnostic: Can be extended through specific thesauruses or ontologies. Verint is present in more than 150 countries with clients in a wide range of industries. The full Impact 360 suite includes:

• Workforce management
• Call recording
• Quality monitoring
• VoC analytics – composed of text analytics, speech analytics EFM
• Desktop and process analytics
• Performance management
• E-learning and coaching

Geography: Global presence.

Sales Model: Each speech and text analytics product can be deployed separately either on-premises(one-time fee) or via SaaS (pricing based on volume).

Table 29. Verint Systems Capabilities

Verint Systems
Information Extraction	Yes
Extraction Methods	APIs Database extractions Reports Completely open analysis scheme Collaborate module Microsoft Office suite NLP Custom dictionaries and keywords Entities, phrases and sentiment extraction
Sentiment Analysis or Opinion Mining	Yes
Big Data Capabilities	Yes
Multiple Language Support (Number of Languages)	Yes (9)
Workflow	Yes
Document Summarization	Yes
Organizing and Structuring
Categorization	Yes
Clustering	Yes
Concept Relevance	Yes
Question Answering	No
Export to Semantic Web Format	No
Information Visualization	Yes
Sector Adaptation
Out-of-the-Box	Yes
Required Third Party	No
Development Tools or Standard Interface for Application Integration	Yes
Real-Time Capabilities	Yes
Predictive Analytics Integration	Yes
BI Platforms Integration
XML Feeds	Yes
REST API	Yes
Special Connector	Yes
Web Services	Yes
Subscription-Based SaaS or Cloud Option	Yes
Open Source	No
Claimed Differentiators	Text analytics solution is part of the company's broader VoC analytics solution and its enterprise workforce optimization suite Root-cause analysis quickly identifies feedback drivers Comparative analysis reporting compares products, competitors, brands, regions, stores, timeframes or any other segmentation of data in a simple two-step wizard Real-time alerts based upon category thresholds, trend and spike reports Satisfaction scoring enhancements support input calculations based on current satisfaction scoring methodologies
BI = business intelligence; NLP = natural language processing; REST = representational state transfer; SaaS = software as a service; VoC = voice of the customer; XML = extensible markup language

Source: Gartner (September 2012)

We based this report on a survey of 55 vendors conducted in April 2012. We include only vendors that have distinct text analytics offerings, not those whose text analytics technology is part of another product.

Note 1. Gartner's Definitions of Text Analytics

For the purpose of this research, Gartner asked the featured vendors to respond to a short RFI about their out-of-the-box text analytics features (see definitions below). The responses were then checked and aligned with Gartner opinion.

Information Extraction

An automated and consistent process for extracting text in different formats from a wide variety of internal and external data sources including email, news feeds, Twitter, Facebook, blogs and other social media.

Extraction Method
An ability to use specialized dictionaries, taxonomies, ontologies, or extraction rules.

Sentiment Analysis
Deep sentiment/emotion/opinion extraction through the computational study of opinions, sentiments and emotions expressed in text. An opinion on a feature is a positive or negative view, attitude, emotion or appraisal from an opinion holder.

Big Data Capabilities
Ability to support technology to handle "big data" e.g., via Hadoop/MapReduce.

Multiple Languages Support
Ability to handle multiple languages within a single document in a text analytics cycle.

Workflow
Ability to create custom workflows or to create or change topics/categories yourself.

Document Summarization
Technologies that can make a coherent summary, of any kind of text, need to take into account several variables such as length, writing style and syntax to make a useful summary to optimize the relevancy of search results.

Organizing and Structuring
Categorization
Content based, request oriented or mixed classification in either supervised document classification or unsupervised document classification.

Clustering
Unsupervised document organization, automatic topic extraction and fast information retrieval or filtering based on algorithms, such as K-means, hierarchical based, graph based, ontology supported, order sensitive.

Concept Relevance
Connect related documents by identifying their shared concepts, helping users find information they perhaps would not have found through traditional search method.

Question Answering
Ability to answer questions posed in natural language in text format.

Export to Semantic Web Format
Export the analytic result dataset in Semantic-Web-friendly format (RDF, OWL, microformats) for data discovery or publishing.

Information Visualization
Manipulation of information from the text analytics result datasets. This will include tasks like adding, removing or rearranging dimensions or metrics, sorting and filtering data, selecting to drill or excluding from display and other analytic-oriented operations.

Sector Adaptation
Provide functions, features and a user interface which can easily be adapted to vertical industries like hospitality, insurance, retail, healthcare, communicationsand financial services.

Development Tools
An API, standard interface to support embedding text analytics functions in applications to offer real-time operation.

Real-Time Capabilities
The ability to capture and conduct text analytics from continuous feeds of text data and provide real-time information to users.

Predictive Analytics Integration
The ability to feed information to predictive analytic applications.

BI Platforms Integration
Delivery of dataset from text analytics result, through the connection to their metadata with specific plug-in or connectors, for reports, dashboard or data visualization tool managed and executed on external vendors' BI platform(s).

Subscription-Based SaaS or Cloud Option
Support software as a service or cloud deployment model based on a subscription model.

Open Source
Is the product distributed under an open source license?

Note 2. Vendors That Failed to Return a Completed RFI Survey

Mega Vendors:

• Oracle (www.oracle.com)

ECM and Search Platform Vendors

• Colbenson (www.colbenson.com)
• Intrafind (www.intrafind.com)
• Microsoft Fast (www.microsoft.com)

PurePlay Platform Vendors

• AeroText/Rocket (www.rocketsoftware.com/products/aerotext)
• Basis Technology Rosette (www.basistech.com/products)
• Daedalus (www.daedalus.es)
• Gate (gate.ac.uk)
• Language Computer Corporation (languagecomputer.com)
• Megaputer Intelligence (www.megaputer.com)
• MeshLabs (www.meshlabsinc.com)
• SRA NetOwl (www.sra.com/netowl/textminer)
• Statsoft (www.statsoft.com/)

Application Specialist Vendors

• Attenex www.ftitechnology.com
• ClearForest/Thomson Reuters (www.clearforest.com)
• Langumatics (www.linguamatics.com)
• General Sentiment (www.generalsentiment.com)
• IxReveal (www.ixreveal.com)
• Marklogic (www.marklogic.com)
• Medallia (www.medallia.com)
• Motivequest (www.motivequest.com)
• OPenAmplify (www.openamplify.com)
• Serendio (www.serendio.com)
• TeraData (www.teradata.com)
• Textkernel (www.textkernel.com)
• ZyLAB (www.zylab.net)

Source: Gartner Research G00232642, Daniel Yuen, Hanns Koehler-Kruener, 14 September 2012