Ensemble method for Text Classification in medicine with multiple rare classes

The paper presents an ensemble method for text classification in the presence of multiple rare classes in the context of medical record data. Specifically, our study aims to classify clinical notes into multiple disease categories, including rare diseases. The Ensemble method involves combining the predictions of multiple machine learning models to predict the patient's diagnosis more accurately. We used three different machine learning algorithms, namely Support Vector Machine, Random Forest, and Naive Bayes, to generate three distinct models and combine their predictions through an ensemble method. The results demonstrate that the ensemble method improves the classification performance compared to individual models. We evaluated this approach on a dataset of 50,000 clinical notes with multiple rare classes.