README

This repository contains the codes used in the Yandex personalized web search challenge. 
Raw log containing query and click data of users spanning for 27 days and similar data for current session for the next 3 days were given using which personalization of search results was to be done - more specifically re-ranking of the 10 URLs shown to the user. More can be found at https://www.kaggle.com/c/yandex-personalized-web-search-challenge/

-> FeatureExtractor is the code to parse the given raw data in log(16 GB), compute the features and prepare output files.
-> DBPopulator is the code to parse the given files and make a DB from it. It was essential for low memory machine(at least for feature extraction). Though with 64 Gb machine, it was not used.