A MACHINE LEARNING MODEL FOR SOBRIETY AND RELAPSE ANALYSIS IN DRUG REHABILITATION

Abstract

Drug abuse has become so paramount among members of society. Although, the initial decision to take drugs is typically voluntary among victims. As drugs are constantly been used, the ability to exert self-control on them is relatively impaired. Thus, abuse is witnessed in different age groups, gender, and celebrities from all walks of life. While addicts become such owing to several factors of curiosity and peer pressure, recreational purpose, source of inspiration, and more, the effect of these drugs can lead to depression, brain stimulation, and hallucination. Managing drug abuse through behavioral or pharmacological means is intended to help addicts stop habitual drug use. Oftentimes, rehab is not effective because the desired change is absent while a proper technological-driven approach to track sobriety and relapse in compulsive drug seeking and usage is also missing. Consequently, in this research, user-friendly and interactive sobriety and relapse predictive management application is developed. Here, addicts' behavioral and demographics with major relapse monitoring factors were clustered to predict the likelihood of relapse. The relapse predictive system using cognitive behavioral patterns, adopts the logistic regression algorithm of the Bayesian network for both training and testing. The essence is to ascertain users’ addiction level, analyze and track sobriety and relapse in order to uncover drug addiction patterns, discover the probability of relapse occurrence towards recommending sustainable rehabilitation decision support