A predictive model that can determine, given a medical abstract, which of 5 classes it falls in.
Medical abstracts describe the current conditions of a patient. Doctors routinely scan dozens or hundreds of abstracts each day as they do their rounds in a hospital and must quickly pick up on the salient information pointing to the patient’s malady. The dataset contains abstracts from 5 different conditions: digestive system diseases, cardiovascular diseases, neoplasms, nervous system diseases, and general pathological conditions. The predictive model can be used to design assistive technology that can identify, with high precision, the class of problems described in the abstract.