why is crab's UserBasedRecommender so slow?
Opened this issue · 6 comments
the benchmark for crab is (http://www.slideshare.net/marcelcaraciolo/crab-a-python-framework-for-building-recommender-systems page-37)
Benchmarks Pure Python w/ Python w/ Scipy Dataset dicts and NumpyMovieLens 100k 15.32 s 9.56 s http://www.grouplens.org/node/73 Old Crab New Crab
however, in my case, I need to take more than 30 minutes to do it, I don't know the reason
my code is
model = MatrixPreferenceDataModel(recommend_data.data)
similarity = UserSimilarity(model, pearson_correlation)
recommender = UserBasedRecommender(model, similarity, with_preference=True)
recommender.recommend("6")
my data is NumpyMovieLens 100k, which contains 100,000 ratings from 1000 users on 1700 movies.
my code is very slow too, it is a problem
my training data:
4701 users, 2721 movies
Yes it is slow , maybe should we build it in the background and save the trained data into an indexer to retrieve them easily later on.
Maybe there is another solution , or did you find anything usefull ?
I have same problem! crab recommendation is very slow.
In my case, it takes 3hours for recommender.recommend(1) by using movelens100k.
What is the problem???
My system is python 2.7 and crab in ubuntu.
How can I reduce the recommedation time?
I guess the project is dead! any alternative please ?
3 hours for recommending on the movielens100k dataset is hard to imagine. If you are looking for basic SVD approaches, you can check the code for the SVD recommender in reco. It takes 15-25 seconds as I have tested.
recommender = UserBasedRecommender(model, similarity, with_preference=True)
recommender.recommend(5)
After running these lines am getting output as : array([], dtype=float64)
what's wrong here. even I copied whole code from http://muricoca.github.io/crab/tutorial.html