Hello, We have an interesting and we guess common machine learning problem.
Polyvore (the company I work for) has a few million products in its index. For each product we have the following information: - keywords extracted from title, description, tags - other meta information like colors, brand, site, price, etc... - user information, for example which users saved this product - CTR data from interaction of users with products on our site. We want to build a recommendation system on top of this data. We are considering using Solr or Mahout (on EMR) to implement this system. If this is a project that you might be interested in helping us with, please drop me a line (guyuan+j...@gmail.com <guyuan%2bj...@gmail.com>). We have a $10K budget for the initial implementation. Thanks, Guangwei