Predicting the Symptoms of Bipolar Disorder in Patients using Machine Learning

Abstract

The modern and fast paced lifestyle in today’s real world lead to high prevalence of mental and psychological disorders like stress, Anxiety and depression in people around us worldwide. The disorder is a result of mood swings and occurrence of oscillations in person’s mind in two states-mania and depression. A complex brain disorder that have affected millions of people across the world is Bipolar Disorder. These conditions led to increase mental health precautions and care using Machine Learning Techniques(ML) for diagnosis and treatment of disease. Using ML, we study patterns in human behavior regularly, identify their symptoms and risk factors to develop a prediction modal. Dataset is visualized to extract meaningful predictions and optimizing therapies. The paper presents commonly used ML Algorithms like Logistic Regression, Support Vector Machine, K-Nearest Neighbor, Naïve Bayes and Decision Trees to study their properties and performance that act as a guide to select the appropriate modal. These modal can bridge the gap between Therapist and patients to revel their problems and embarrassment to expose their illness. This is the key task in selecting the features from dataset and applying the appropriate modal.