Prediction Model for Amphetamine Behaviors Based on Bayes Network Classifier
Kumnung Vongprechakorn, Narumol Chumuang, Adil Farooq · 2019
This paper focus to present a prediction model for drug addiction of the accused in type 1 drug abuse case as amphetamine. Case studies in the Suan Phueng police station area Ratchaburi province. The data set that used for modeling is obtained from the collection from Suan Phueng police station from 2016-2018. The data set 1,598 items consist of gender, age, number of offenses, education status, nationality, occupation, and non-drug abuse. For our contribute a prediction model into two classes as “take” and “untake” by using data mining techniques namely Bayes Network classifier. In our experimental, a group of bayes are used to comparison such as Bayes Network, Naive Bayes and Naive Bayes Updateable. The results displayed the Bayes Network classifier shown the highest accuracy rate, Naive Bayes and Naive Bayes Updateable with 81.53, 80.85 and 80.85 respectively.