Exploiting textual and numerical data to improve predictive models for depression

In this project, we improve methods to interpret free text in a large dataset from the domain of mental depression using sentiment analysis techniques.