Stacked SVM model for Dysthymia prediction in undergraduates students

Dysthyrnia has been the most prevalent form of clinical depression among all mental and physiological illnesses. It affects a large number of undergraduate students worldwide, and is a matter of concern for clinical psychologists. DSM-5 based personalized diagnosis has been an indispensable technique since long and is exclusively based on subjective behavioral interpretation of a clinical psychologist/psychiatrist. However, such individual behavioural assessments can be sensitive to variance, slowness and experience of the interviewer. Precise and authentic diagnosis of dysthyrnia can facilitate an appropriate psychological response programme and in general increase the patient's psychological well-being. Hence, machine learning based analysis of behavioral data is one such quantitative approach which can expedite in the early diagnosis of dysthyrnia. As a result, the goal of this work is to develop a stacked support vector machine (SVM) based classification model for the preliminary screening of dysthyrnia in undergraduates. In the present work, behavioural data was collected from 98 undergraduates, and simulations were performed, and experimental findings are recorded. In comparison to single kernel SVM and other standard classifiers, it has been observed that the proposed stacked SVM model achieves an accuracy of 85.74%.