Using Data Mining Techniques for Predicting Alcohol Consumption in Portuguese secondary schools

Many factors are involved in student academic failure. It seems academic failure rate is high in Portuguese secondary school students. Although many factors such as family situation, parent’s occupation, student’s study time, and many other factors are effective, but one of the determining factors is alcohol consumption among Portuguese secondary school students. Data mining techniques can be used to discover knowledge. Classification is a data mining technique that assigns items in a collection to target categories or classes. In this paper we predict the amount of alcohol consumption in two Portuguese secondary schools using five level classification methods. We investigate the effect of various factors on consuming alcohol by measuring the correlation between different factors. Also the distribution of alcohol consumption levels among both genders and the effect of the mother’s job on the consumption level were investigated.

کلیدواژه ها:

artificial neural network ، classification ، data mining ، decision tree ، K-nearest neighbor ، Naïve Bayes ، prediction