Boston, June 28, 2006 - Yellow Pages Group, Canada’s largest directory publisher, has released a cutting edge local search website powered by ApproxiMatch that will serve to consolidate the Company’s leadership position in the online directory market.

ApproxiMatch fuzzy search methodology and machine learning features are uniquely suitable for local search, and help set the YellowPages.ca™ online directory apart. The beta version of the YellowPages.ca™ site can be found at http://beta.yellowpages.ca.

"People always associate Google with search, but local search is a very different problem and it require a very different solution," added ApproxiMatch founder and CEO, Erez Shechter. "Whereas Google was built to solve information retrieval on the web, the ApproxiMatch search engine was designed to provide the most relevant local results, based on both structured and unstructured data."

Also being deployed by Verizon SuperPages, the largest U.S. Internet Yellow Pages site, ApproxiMatch’s fuzzy search enables users to pose queries with multiple keywords, brands or phone numbers. A location search box enables users to search by landmark, region, postal code. Users can even misspell words in their searches, as ApproxiMatch technology automatically detects and corrects spelling errors. Searches can also be refined by brand, payment method and hours of operation.

Fuzzy search is based on fuzzy logic formalism, which mimics the human ability to rapidly select the most relevant results from a large data set. Results are selected and ranked on the basis of real-time contextual matching rather than repeated application of exact matching which is common to all other search engines.