Using Naïve Bayes Algorithm to Predict and Classify Alcohol Addiction Severity: A Machine Learning Approach for Public Health Interventions

Alcohol addiction has increasingly emerged as a significant concern in global health, with current methods of prediction and classification revealing certain limitations. The principal objective of this study was to deepen the understanding of predicting and classifying alcohol addiction levels by employing the Naïve Bayes Algorithm and K-means Clustering. Through a thorough survey, data from 500 participants were collected, shedding light on factors such as the frequency of alcohol consumption and associated negative impacts. The methodology utilized the Naïve Bayes Algorithm, registering a notable accuracy of 95%, precision of 93%, recall of 97%, and an F1 Score of 95%. Concurrently, the K-means Clustering method effectively delineated three distinct levels of addiction: less addicted, mildly addicted, and highly addicted. When juxtaposed with existing literature and methodologies, the study's approach showcases superior accuracy and a refined classification system, offering a potent tool for healthcare practitioners to identify and address alcohol addiction. Potential avenues for future exploration include integrating varied algorithms and probing into other facets of addiction.