Due to the high cost of developing novel drugs, drug repositioning represents a promising alternative method of treatment. In this project, we use a network medicine approach to identify novel drug candidates for repositioning. Using a powerful machine learning text-mining application, we built knowledge networks containing connections between human diseases and genes or drugs mentioned in the scientific literature published in the past decades. This approach reveals several drugs that target key disease-related genes, which have never been used to treat these disorders to date.