Main Content

Semantic Search and Recommendation with Python and Solr

A talk by John Berryman at PyTennessee

About the Talk

February 22, 2014 5:00 AM



Built upon Lucene, Solr provides fast and highly scalable full-text search capabilities. However, under the hood, Solr is really just a sophisticated token-matching engine. What's missing? Semantic search and search-aware recommendation.

Semantic-search allows Solr to infer the meaning of the terms that users are searching for. It does this by processing the documents and identifying commonly co-occurring terms. So, for instance, if you're looking for documents about "software architecture" then you might also be interested to find "programming design pattern" documents as well even if they do not contain the term "software" or "architecture".

Search-aware recommendation takes this concept to the next level. By comparing user similarity based upon products previously purchased, it is possible to make search-aware recommendations which will be perfectly tailored to the user's expected purchasing behavior. For instance, some customers will search for "cheese" and think 5 year Gouda wrapped in oak leaves, while others just want their Velveeta thin slices!

In this fast-paced discussion I will describe the basic mechanics of search and outline the need for semantic search and recommendation. Then, I will provide a demonstration of how the basic mechanisms of search can be augmented so that both semantic search and recommendation are possible.

Ratings and Recommendations

This Talk hasn't been rated yet. Sign In to rate Talks.

comments powered by Disqus