Project Title: Development of machine learning algorithms for the early diagnosis of delinquent behavior in juveniles.
Juvenile delinquency is the habit of committing criminal offences by an adolescent, (or juveniles, i.e. individuals with age at which ordinary criminal prosecution is possible). Clinically persistent pattern of antisocial behaviour in which the individual repeatedly breaks social rules and carries out aggressive acts that upset other people is known as conduct disorder. During the beginning of adolescence, children develop more advanced patterns of reasoning. Hence, there is a higher risk of development of deviance behavior in juveniles. As deviant behavior usually develops during adolescence, it is very much essential for the early screening of such individuals. Human screening of such individuals by a psychologist is quite difficult in many cases as several individual, parental, and environmental parameters are needed to be correlated for final decision making. Thus, some efficient computer aided mechanism is required to be developed for the early screening of adolescents with suspicious disruptive behavior. Those identified adolescents with disruptive behavior must be provided with proper psychological counseling to bring them back to normal behaviors. Taking this problem into consideration, an effort was be made to develop an automated system for the early screening of delinquent behavior in adolescents. In this project we focus on using biological and socio demographic factors along with machine learning principle for the early prediction of delinquency in adolescents. A standard psychometric questionnaire (DSM-5) is used for collecting the above risk factors to assess the current psychological condition of the child. These set of weighted feature scores will be fed to the machine learning module (clustering algorithm) for the categorization of an adolescent behavior into two classes i.e. normal and delinquent. Additionally, Child Behavior Checklist (CBCL) scale is used to collect features related to severity of delinquency. With respect to these features the developed machine learning model will categorize the adolescent based on the level of risk of showing delinquent behavior i.e. low, moderate, and severe risk.