Identification of drug-target interactions via multi-view graph regularized link propagation model.

Diseases are usually caused by body's own defects protein or the functional structure of viral proteins. Effective drugs can be combined with these proteins well and remove original functions to achieve the therapeutic effect. The biochemical approaches of drug-target interactions (DTIs) determination is expensive and time-consuming. Therefnal-based methods have been proposed to predict new DTIs. In order to solve the problem of multiple information fusion, we propose a multi-view graph regularized link propagation model (MvGRLP) to predict new DTIs. Multi-view learning could use the complementary and correlated information between different views (features). Compared with existing models, our method achieves comparable and best results on four benchmark datasets. [ABSTRACT FROM AUTHOR]