Strange predictions using SVD in mahout

Question

I'm trying to build svdrecommender using mahout. Code is simple:

DataModel model = new FileDataModel(new File("C:\\data.csv"));
SVDRecommender recommender = new SVDRecommender(model, new SVDPlusPlusFactorizer(model, 10, 20));

All my ratings are doubles between 0 and 1. However recommender in most cases predicts values above 1. How could it happen? Is it a feature of svd algorithm?

Answer 1

SVDRecommender uses approximate decomposition of ratings' matrix into two other matrixes. So their product can contain arbitrary numbers in cells.