Sign In

Communications of the ACM

ACM TechNews

Fast Algorithm Extracts, Compares Document Meaning

View as: Print Mobile App Share:
archive binders


Technical University of Wroclaw researchers have developed an algorithm that reduces the complexity involved in comparing two documents by merging a computationally efficient statistical approach to text analysis with a semantic component.

The researchers say the algorithm could revolutionize the way archived documents are searched, and enable knowledge to be extracted far more readily than with conventional tools. They note the approach also solves three major problems faced by most users of today's search engines, including the lack of familiarity with the advanced search options of search engines, and the strict nature of the options that are unable to catch nuances in user information needs. The algorithm also addresses the difficulty and effort necessary to type a long query.

Wroclaw's Andrzej Sieminski says the key to their work is using statistical similarity measures to assess semantic similarity.

View Full Article

Abstracts Copyright © 2012 Information Inc., Bethesda, Maryland, USA 


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account