An Improving efficiency of data classification for notebook computer purchase decisions of Nakhon Pathom Rajabhat University students using data balancing

Abstract

The purpose of this research is to balance the data and compare the efficiency of the model for predicting notebook computer purchase decisions by using 4 techniques: Decision Tree(J48), Naïve Bayes, k-nearest neighbors' algorithm (k- NN), and multi-layer perceptron. These technique are used to compare the performance of the appropriate classification model for predicting notebook computer purchases using the collected data of 101 Nakhon Pathom Rajabhat University students and the Weka software version 3.8.6. Due to the data is imbalanced, the accuracy is very low, it is not suitable for use. The researcher, therefore, solved the problem by using additional data synthesis methods to increase efficiency. The results are satisfactory. It is found that the K-NN technique is the most suitable model used to classify the purchasing decisions of notebook computers of Nakhon Pathom Rajabhat University students. Get the highest accuracy of 80%, a precision of 0.79, and a recall of 0.80.