You have users, services, and vendors. You should decide what you want to recommend. Service? Vendor? Service of Vendor?
Assuming the latter combine the services and vendors into a single ID space: vendor1-service1, vendor1-service2 … Then decide what method you want to create recs. We are generally recommending you use Hadoop “itemsimilarity" or "spark-itemsimilarity" jobs to create an indicator matrix and use a search engine to query for recs. But you could also use the Hadoop-based recommender from Mahout. Input to the Hadoop Mapreduce jobs will take input like this: user, item 0,0 0,10 your recs will be returned using the same integer IDs so you will have to translate your “user1” and “vendor1-service1” into non-negative contiguous integers If you use spark-itemsimilarity you can use your string IDs user, item user1,vendor1-service1 user1000,vendor10-service1 ... To use a search engine have a look at this short book, which describes the process: https://www.mapr.com/practical-machine-learning On Sep 29, 2014, at 9:53 AM, vinayakb malagatti <[email protected]> wrote: Hi all, I have table something looks like in DB : rating table <https://docs.google.com/spreadsheets/d/1PrShX7X70PqnfIQg0Dfv6mIHtX1k7KSZHTBfTPMv_Do/edit?usp=drive_web> Thanks and Regards, Vinayak B
