Automated Brain Tumor Prediction System using Natural Language Processing (NLP)

Abstract

In this paper, we proposed an Automated Brain Tumor Prediction System which predicts Brain Tumor through symptoms in several diseases using Natural Language Processing (NLP). Term Frequency Inverse Document Frequency (TF-IDF) is used for calculating term weighting of terms on different disease’s symptoms. Cosine Similarity Measure and Euclidean Distance are used for calculating angular and linear distance respectively between diseases and symptoms for getting ranking of the Brain Tumor in the ranked diseases. A novel mathematical strategy is used here for predicting chance of Brain Tumor through symptoms in several diseases. According to the proposed novel mathematical strategy, the chance of the Brain Tumor is proportional to the obtained similarity value of the Brain Tumor when symptoms are queried and inversely proportional to the rank of the Brain Tumor in several diseases and the maximum similarity value of the Brain Tumor, where all symptoms of Brain Tumor are present.