Using Engine Scoring and Predictive Analytics to Boost Search Accuracy
A Video Recap of My Speaking Engagement at Lucene/Solr Revolution 2015
Many of my recent discussions have revolved around search engine scoring. With big data, it is now possible to harvest user event data, such as search logs and click logs, for the purpose of computing user-based search accuracy metrics. This process can really make continuous improvements to search engines, and our team is working to enhance the algorithms everyday (check out my blog from the trenches).
In the Fall, I had an opportunity to lead a session on this very same, fascinating topic at Lucene/Solr Revolution 2015. During my Search Accuracy Metrics & Predictive Analytics presentation, I discussed the algorithms and processes we use for calculating search accuracy metrics. These metrics are enormously helpful for computing and comparing search engines' accuracy before the engine is deployed.
We also use big data to extend the methods to include predictive analytics for business metrics (conversion rates, abandonment rates, etc), machine learning for relevancy optimization, and incorporating optimized relevancy formulae into Lucene-based or other search engines.
I invite you to watch the session recording below, with demo and examples, to get a deep-dive into this topic.
Wondering how you can make the most of the advanced query processing techniques mentioned in this presentation? View The Magic and Wonder of Query Parsing video-blog for guidelines and best practices.