Proposing alternative medical diagnoses

Leveraging methods from Machine Learning to mine information found in electronic health records.

Electronic health records (EHRs) of patients are usually stored on a central server at hospitals. These are a possible gold-mine of information that is currently large under-utilised. In this project, we aim to make use of modern Machine Learning, Deep Learning and Natural Language Processing methods to extract useful information from these records, which can help medical doctors to make better diagnoses. The probabilistic model makes use of the information regarding the current patient and matches it to other information about other patients in the database. The model is then used to list a number of possible diagnoses, which can be useful in helping medical doctors in their work or for planning and resource allocation.

The image is used under Creative Commons with credits to NEC Corporation of America at FlickR.