RecommenderDemo
RecommenderDemo copied to clipboard
large dataset
first is to congratulate on the nice code .anyway I would like to use your code to create an advert recommendation system for a website am planning to build. I project to the tune of three million users and many transactions which involve computations. is the code suitable to handle such data? please advice
Hi Bill,
Please go ahead and use what you can. Three million users, you'd be looking in the region of 200M ratings. At this point I'd be looking to scale with Hadoop to split the work out.
There's a nice walkthrough here (not by me): http://chimpler.wordpress.com/2013/02/20/playing-with-the-mahout-recommendation-engine-on-a-hadoop-cluster/