Publication

Abstract : Psychoactive drugs, which influence mood, awareness, thoughts, and behavior, are widely researched upon due to their impact on both individuals and society. This work presents a Psychoactive Fusion Machine Learning framework consisting of a classification model to identify psychoactive drugs based on chemical attributes and drug usage, utilizing personality traits and demographic data. Combined with a regression model to predict impulsivity and sensation-seeking scores (ImpSS) using demographic and personality features. This model leverages various algorithms including XGBoost resulting in an accuracy of 85 % for the identification of psychoactive drugs, Logistic Regression and Support Vector Machine yielding high accuracies of around 97 % when used in the prediction of behavioral patterns under drug influence, Random Forest with an accuracy of 83 % in the prediction of Impulsivity. The results demonstrate the potential of combining personality psychology and chemical analysis with machine learning to achieve precise and explainable predictions, opening pathways for applications in healthcare, policy-making, and substance use research, such as early predictions of critical care patients and patients who need an immediate intervention in order to avoid the severities of the abuse.