Betel Nut Addiction Detection Using Machine Learning

Abstract

Betel nut addiction has become a psychoactive addiction to hundreds and millions of people around the world. Among the four common hallucinatory addictions in the world, betel nut addiction is one of them after tobacco, alcohol, and caffeine. It is more acute in the countries like Bangladesh, India, Pakistan and Taiwan. This ancient habit is much harmful to the human body as it contains substances like arecoline, which is more similar to nicotine. Consuming it on a repetitive basis causes the deformation of dental structure and change in the oral mucosa. Also, people have a risk of facing several diseases such as oral cancer, neck, and head cancer, submucous fibrosis, etc. by consuming it. Unfortunately, in Bangladesh, rural women are more addicted to this harmful practice. The purpose of this research is to detect betel nut addiction so that its disadvantages can be spread among the people. In this paper, the machine learning algorithms detect whether a person is addicted or not addicted to betel nut. Nowadays, machine learning is the most popular tool for using algorithms that compute data and learn from it for decision making or prediction about something. Thus, machine learning approach was chosen. The methodology has required data, which was collected manually by analyzing some factors from previous research papers on betel nut addiction. The k-nearest neighbor (kNN), Support Vector Machine (SVM), decision tree, logistic regression, random forest and Naive Bayes algorithm was used for classification and machine learning technique have been performed and evaluated them regarding performance matrices with a featured dataset. Among all the algorithms, random forest performs the best with the highest accuracy of about 99.0% with less training time.