Modern organizations have access to abundant sources of data that contain jewels of useful knowledge. Emails, web pages, memos, call center transcripts, survey responses, claims notes, legal cases, patent descriptions, research articles, and incident reports – all hold valuable pieces of knowledge that are hard to discover because they are hidden within large volumes of raw data or structured datasets not typically indexed for key term searching. Overwhelmed by the glut and variety of data, law firms and corporations are seeking new methods to analyze large volumes of structured and unstructured data to not only discover key knowledge, but to create case or review strategy as well.
Automated Knowledge Discovery & E-Discovery Text Analytics
Automated Knowledge Discovery ("AKD") is our service that derives knowledge from text and structured data, including XML files, databases, spreadsheets, web sites, documents, and emails using Natural Language Processing, Machine Learning, Interactive Visualization and Report Generation. AKD enables you to gain valuable insight into large diverse data sets by leveraging intelligent data mining and analysis tools to undertake a range of knowledge discovery tasks, including:
Data Source Integration
- XML Files
- Email / PST Files / Exchange Datastores
- File Shares
- Delimited Files
- RSS Feeds
- Web Sites
Distinct Text Identification
Compares records to determine whether they are duplicates, though comparison is fuzzy in the sense that the two values do not have to be 100% identical in content and form. Similar to "near duplicate indentication".
Key Term Linking
Generates a visualization of associations between various keywords so that that relationships between words are well-represented in a set of documents or a column of text values.
Utilizes phrases find and identify similar groups of documents or records based on the contents of text.
- Human names
- Money Amount
- Geographic Names
Determines the important concepts discussed in a text via the identification the most frequent keywords and the ability to drill down on each keywords to see how the words are being mentioned.
Statistical determination of phrases by examining the co-occurrences of consecutive words within the text. Stated more simply, if two words occur next to each other repeatedly in several sentences across several documents, it can be statistically assumed that these words constitute a phrase.
Displays correlations and clusters of terms extracted from a set of documents or text values with the ability to drill down to review the specific relationships.
Searches for phrases, proximity, by sentence or paragraph, synonyms and alternate term forms and enables the ability to see a ranked list of matching records, search within the results of previous searches, refine search query by providing positive and negative document-based relevance feedback, and the highlighting of occurrences of words in documents for easy navigation.
Our Automated Knowledge Discovery Services provide valuable insight to any collection of discovery data, far beyond key term browsing typically associated with litigation support systems, many of which simply do not have the ability to process structured data such as SQL databases or "as found" versions of email stores. Additional benefits of AKD Services include:
- Mobile platform deployable worldwide
- Fee structure based on time, not volume
- No long term commitments, user fees or storage fees
- Compliments, not replaces, current review workflows and systems
- Ideal platform for data sampling