Neural Network Based System for Disease Prediction

In the medical field, among the most important aspects are considered the good interpretations of data and the establishment of correct diagnoses. In this context, a better prediction taking, into account the patient's health problems, can lead to an appropriate treatment in a shorter time. The paper presents a classification system based on a feed-forward neural network to predict the type of some severe diseases based on known symptoms. The proposed system is learned and tested on a public dataset where 41 diseases were considered. The validation of the proposed method was done in terms of performance metrics in two scenarios with different proportions of data, and they were compared with similar papers from references.