Text Analytics

Make Decisions Based on a Complete Understanding

Attivio uses both directed and undirected processes to analyze text at scale, including tokenization, lemmatization, clustering, key phrase and entity extraction, classification, and rules-based tagging.

While many organizations struggle to get their arms around just structured data - leaving the vast majority of their unstructured content largely untapped - Attivio empowers information workers at leading enterprises to harness the full value of all their information. By automatically identifying key concepts, extracting entities, and analyzing sentiment – all with multi-language support – Attivio reveals the untapped business value typically hidden in unstructured text and unifies it with the related structured data to present a complete view.

Entity Extraction

Attivio detects and indexes multiple entities, such as people, companies, and locations.

FEATURES INCLUDE:

  • Entities can be based on supplied, supplemental dictionaries, on patterns, or rules
  • Entity extraction ships with the standard product

Concept Extraction

Attivio uses statistical techniques (e.g. machine learning) to extract important phrases.

FEATURES INCLUDE:

  • Key phrases to classify information
  • Facets to facilitate information discovery
  • Filters to refine user queries

Sentiment Analysis

Stay on the pulse and be actively aware of stakeholder perceptions about your brand, products or services. Better understand your competition to improve business processes and the overall customer experience.

FEATURES INCLUDE:

  • Leveraging Attivio’s Classification Engine, detect positive and negative sentiment from any source: internal (email, live chat, online surveys, call center notes) or external (social media, news, blogs, online communities)
  • Patented technology to calculate sentiment at an entity level (person, product, etc.) and/or overall document level
  • Graphically analyze changes in sentiment over time
  • Review and refine sentiment models using detailed explanations of sentiment scoring results