Prediction of Graduation with Naïve Bayes Algorithm and Principal Component Analysis (PCA) on Time Series Data

Abstract

The percentage of students who graduated on time can be predicted with data mining methods. This research aims to provide earlier information regarding students who are at risk of not graduating on time. Thus, the study program can take appropriate action before it is too late. Several classification methods can be used for prediction. Our research combines Naïve Bayes with Principal Component Analysis (PCA). PCA is used to simplify complex academic data. The PCA result has a more straightforward structure to be processed using Naive Bayes classification. This research uses four batches of student academic performance data in Informatics Study Program, Telkom University. The dataset is partitioned by academic year to obtain time-series data of each student. The combination of PCA and Naïve Bayes algorithms obtained better results than classification using Naïve Bayes only, with 6.04% higher accuracy on average.