Lucene/Solr Revolution 2017 has ended
View analytic
Thursday, September 14 • 1:30pm - 2:10pm
Personalized Search Results and Job Recommendations at Dice.com

One drawback of machine learned ranking models is the lack of personalization; the algorithm does not typically make use of a user’s behavior to tailor search results to the user’s own preferences, instead relying on relevancy judgements usually collected from a small subset of users. Relevance feedback takes a different approach – use either implicit or explicit feedback from the user to improve relevancy. By monitoring a user’s individual search behavior, their prior search and browse behavior, or by asking them for explicit feedback on the relevancy of results, the search engine can better adapt search results to the individual user, allowing for a more personalized search experience. Another form of relevance feedback, so called “Blind Feedback”, uses the initial set of search results to expand and re-execute the original query to improve recall, without the need for explicit synonym files. A user’s profile can also be used directly to improve search relevancy, or to provide content-based personalized recommendations.

In this talk, we will discuss these various approaches to improving relevancy, and how they can be incorporated into Solr via simple plugins. All code used in the presentation will be made available on GitHub following the presentation.

avatar for Simon Hughes

Simon Hughes

Chief Data Scientist, DHI / Dice.com
Simon is currently the Chief Data Scientist at Dice.com, the technology professional recruiting site. He is also a PhD candidate at DePaul university, studying a PhD in machine learning and natural language processing. At Dice, he has developed multiple recommender engines using Solr... Read More →

Thursday September 14, 2017 1:30pm - 2:10pm
South Seas A