Montreal - June 8, 2005 - Provalis Research is pleased to announce the release of WordStat 5, a content analysis and text-mining tool offering new features that dramatically improve the flexibility of text categorization and document classification processes and that enhance support for knowledge-discovery tasks. WordStat 5 delivers improved performances and more than 40 new features, including:

• Rule-based codification of terms using Boolean (AND, OR, NOT) and proximity (NEAR, BEFORE, AFTER) operators.
• Ability to defined external text preprocessing modules.
• Automatic document classification using Naive Bayes and k-Nearest Neighbor methods.
• Powerful text retrieval features that are fully integrated within tables and visualization tools.
• A vocabulary extractor for identification of technical terms, product names and company names.
• A feature to store normative frequency data to disk and to compare observed frequencies to saved norms.
• New clustering options such as second-order proximity and probabilistic co-occurrence measures.

Provalis Research also announces plans for the future release of a software developer's kit (SDK) for the integration of WordStat categorization and classification technologies into applications, databases or document management systems.

For more information on all the new features introduced in Version 5 or on related products, or to download a free fully functional version of WordStat, visit Provalis Research’s web site at:

http://www.provalisresearch.com

Founded in 1989, Provalis Research provides a family of software systems that is flexible, easy to use and offers an unsurpassed integration of statistical analysis, qualitative data analysis, content analysis and text mining.