Comparisons of the Alcohol Risks Prediction Models with Data Mining Techniques

Abstract

The objectives of this research were 1) to study and develop an alcohol addiction risk forecasting model using data mining techniques 2) to compare the effectiveness of the alcohol addiction risk forecasting model using data mining techniques. Which includes techniques K-Nearest Neighbour technique Naive Bayes and Random Forest techniques by studying a sample of 200 people from a community in Sa Kathiam Subdistrict Nakhon Pathom province, year 2021, data were collected by using a questionnaire consisting of general data and an alcohol drinking behavior assessment form. WEKA was used for analysis to determine Accuracy, Precision, Recall, F-Measure. From the study, it was found that the technique that gave the best analytical efficiency was Random Forest algorithm with an accuracy of 99.50%, a precision of 99%, a recall value of 100%, a f-measure of 99.50%, followed by Naive Bayes. The accuracy is 99%, the precision is 98%, the recall is 100%, the f-measure is 99%, and the K-Nearest Neighbor is 97.50%, the precision is 97.00%, the recall is 98% and the f-measure is 97.50%