Risk Factor Analysis of Diseases Using Machine Learning Techniques

Analysing the risk factors of Mental health from Electronic Health Records is a challenging task as it is difficult to assess the prevalence of diseases due to lack of culturally adapted and validated assessments. In this study, we find the risk factors of Memory deterioration using Machine Learning techniques by implementing Correlation, Regression Analysis and Random Forest algorithms on MYNAH cohort (Mysore Studies of Natal effect on Ageing and Health) which was carried out at the Epidemiological Research Unit, CSI Holdsworth Memorial Hospital, Mysuru, South India. Correlation is used to find the influence of one parameter on the other which play roles in identifying risk factors of Memory deterioration. Regression analysis helps in estimating the relationships among parameters that are used for disease prediction. Random forests or random decision forests algorithm brings extra randomness into the model to search for the best parameter among a random subset of parameters. It is an ensemble learning method for classification, regression and other tasks in which a multitude of decision trees are constructed at training time and the class is output. In Classification problem, the ensemble of simple trees vote for the most popular class. In the Regression problem, the responses are averaged to obtain an estimate of the dependent parameter. Implementation of tree ensembles has lead to significant improvement in prediction accuracy. This work facilitates health care organizations to perform analysis on sector of population prone to various diseases using Electronic Health Records and educate people regarding the risk factors of diseases to enable effective therapy at the right time and place.