Ask Jeeves boosts Q&A database to 300 million

The Ask Jeeves search firm built the pairings using its AnswerFarm technology, which crawls the web to index questions from numerous sources, and uses semantic technology to filter out less meaningful answer formats.

The idea is that a user types a question into the search box, and is presented with web page results displaying the answer. An auto-complete function is included to predict any possible questions and save the user from having to type the entire query.

The firm claimed in an Ask.com blog post to have grown its Q&A database from 100 million to 300 million pairs in a few months.

"Our semantic search technology advances in clustering, rephrasing and answer relevance enable us to determine when we have multiple questions that semantically mean the same thing, so we can aggregate those Q&A pairs, filter out insignificant and less meaningful answer formats, and thus find the most relevant answers," said Ask.com.

A few test queries conducted by us found the site generally good at answering basic questions, and there were helpful suggestions displayed in case the question had not been answered satisfactorily.

However, similar searches in Google also produced fairly accurate results, so it will be interesting to see whether Ask.com can differentiate with this functionality.

The firm will certainly be hoping that the new capabilities help it to improve its position in the market. According to current stats from web data firm Net Applications, Ask.com has a search market share of just under one per cent, behind Google, Yahoo, Microsoft and AOL.