Preventing Measures In Bipolar Disorder Via Machine Learning

Abstract

The technical advancement in recent days has made a drastic increase in the amount of data generated from the physical, cyber, and human world. The collection of data at a huge scale makes sense only if the data is actionable and can also be used for making decisions. Data mining provides needed assistance at this stage by inspecting the relationship in the data and offers needed insights to the data owners. Moreover, the significant insights obtained are shared with third parties for further analysis. In this situation, numerous varieties of information is processed and the integration of machine learning with big data technology has made prominent aspect in the identification of bipolar disorder. It is a complex genetic disorder characterized by episodes of mania and depression. It affects 1% of the population worldwide. It is a major under-addressed public health problem which causes a significant burden on caregivers. High heritability and familial relative risk indicate the role of genetics in the etiology of the disorder. In this research, data is handled using an improved auto encoder and deep neural network. The feature selection is accomplished using ISAE and classification is accomplished using DNN. The proposed approach outperforms the existing state of art techniques.