Unsupervised Discovery of Drug Side-Effects from Heterogeneous Data Sources

摘要: Drug side-effects become a worldwide public health concern, which are the fourth leading cause of death in United States. Pharmaceutical industry has paid tremendous effort to identify drug during development. However, it is impossible and impractical all them. Fortunately, can also be reported on heterogeneous platforms (i.e., data sources), such as FDA Adverse Event Reporting System various online communities. existing supervised semi-supervised approaches not practical annotating labels expensive medical field. In this paper, we propose novel effective unsupervised model Sifter automatically discover side-effects. enhances estimation by learning from measuring platform-level user-level quality simultaneously. way, demonstrates better performance compared with terms correctly identifying Experimental results five real-world datasets show that significantly improve state-of-the-art approaches.