Clay Pryor

Paul Anderson photo

R&D S&E, Computer Science
Sandia National Labs

Clay is a principal member of technical staff at Sandia National Laboratories (Sandia) where he has spent over 30 years contributing to many software development efforts ranging from stand-alone disconnected systems to enterprise web applications. For the past few years he has been given the opportunity to lead Enterprise Search efforts where he has focused on upgrading and enhancing a custom enterprise search application, Solr configurations, custom crawlers, and custom connectors to improve search results delivered to the internal Sandia community. Most recently he has lead the evaluation and deployment of the COTS Lucidworks Fusion software product for enterprise search at Sandia.

Clay Pryor is speaking at the following session/s

Configuring Recommendations for Personalized Search at Sandia National Labs

Wednesday | 1:30PM - 2:10PM |

Learn how Sandia configured personalized search for enterprise search users in days (not weeks, or even years) from data gathering and model building to query configuration. In enterprise search, the assumed preference of each user is the number of times that they have previously clicked on pages (an observed weight) that is then used to "co-cluster" with other users to make predictions about what pages they will find useful. Recommender systems, colloquially RecSys, are a subclass of information filtering systems that seek to predict the rating or preference a user would give to an item. Alternating Least Squares (ALS) matrix factorization is a common technique in collaborative filtering and has been proven to be an effective solution to the co-clustering problem, where the primary model for recommendations must be trained through the “ALS Recommender” job. Explore Lucidworks Fusion out-of-the-box RecSys tools for a multitude of applications.

Audience Takeaway
Learn how to configure personalization for enterprise search using Lucidworks Fusion out of the box with signals.

Intended Audience
Enterprise search managers, engineers, and other staff interested in introducing personalization based on signals into their search results.

All Levels