Factorization vs. Regularization: Fusing Heterogeneous Social Relationships in Top-N Recommendation，很不错的文章，one-model data 适合与regularization，bipartite data 适用于factorization @袁全V @lijiawei166
