Metarank is an open-source Learn-to-Rank personalization tool. It receives a stream of visitor’s interactions, maps it to a set of typical ML features, and reorders your static ES search results to maximize the final metric, like CTR. With off-the-shelf support for common feature extractors like UA, GeoIP, referer parsers, and visitor profiling, you can make your ES search results personalized with no extra code written. Quit changing weights in your hand-crafted ranking formula, and use Metarank to leverage self-tuned ML models like LambdaMART to optimize your ranking!
