A Method to Detect Amphetamine Addiction using Shallow Learning and Deep Learning

Drug addiction is a critical issue, necessitating actions to prevent severe physical and mental health problems. Among all drug addictions, amphetamine addiction rank third in addictive substances according to WHO. In this work, a method has been suggested for amphetamine addiction detection. The input features used to detect the amphetamine addiction are personality features and demographic features. The features are then selected using random forest method. Out of all features, fifteen features are selected except the N-score. The selected features are then given to deep neural network for detecting amphetamine addiction. The sensitivity, specificity and accuracy of the proposed method has been found to be 0.97, 0.86 and 0.92 respectively. The precision, recall and F1-score of the proposed method has been found to be 0.87, 0.97 and 0.92 respectively. The proposed method can be used as an assisting tool for counselling addicted individuals.