Student Posters Repository

Recommendation systems become a much needed tool for predicting user preferences. The development of recommender systems subsume in addition to information retrieval, models and metrics creating, detection of key parameters and many machine learning applications. Model creating is a computationally demanding problem, that can use heterogeneous models as a tool for more effective computational distribution. The poster presents a combination of heterogeneous models and the problem of preference recommendation by using matrix factorization.