Results interpretation

[sunithak]

I have tried running the autoint and svd++ on movielens data that you provided. The results are not very good. The recommendations makes no sense. No user collaboration, items also not very similar. I have tried youtube ranking also with merged dataset. Genres are also very different from each other. What is the base for kpi. Which algorithm can i use to run real time data from your collection.

[massquantity]

The example data only has 100 thousand rows, which is far from enough to train a good model. Besides, are you implying that a "sensible" recommendation is items with similar genres? And what kind of real time data you want to use?

[sunithak]

I am using ecommerce data, let's say fasion industry. Still results are not getting better. Which algorithm has best performance in all of them, so that i can try that one. I am thinking of adding content based algorithm results also to improve results in general. I am using this data: http://jmcauley.ucsd.edu/data/amazon/. Could you please help me build best rec systems? Thank you :)