Risk Prediction of Major Depressive Disorder using Artificial Neural Network

Abstract

Major Depressive Disorder (MDD) is a serious medical condition that can affect many areas of a person's daily life significantly. MDD, caused by a combination of factors, will be debilitating if not detected and managed early. This is why it is the leading cause of disability around the world. If detected early, several treatment and management programs can be done, for example, change of lifestyle. There are models developed to predict the risk of individual suffering MDD but they have low sensitivity and specificity. In this study, a new MDD risk prediction model is developed using a novel equation and Artificial Neural Network (ANN). The model is created using risk factors of MDD that are categorized into three groups, which are psychological, social and biological. Two predictor methods are applied, first, using a conventional equation, then using an ANN tool. From the results, the conventional equation is able to provide the risk estimation for MDD. After comparing, ANN showed the ability to calculate the risk prediction of MDD with 70% test accuracy and found to have a better sensitivity and specificity than the existing models.