A Study and Development of Forecasting Model for The Suitability Characteristics on The Applying Major Selection by Using Data Mining Techniques

Abstract

The purposes of this research were to: 1) study the model of qualification prediction on applying major selected by using data mining techniques; 2) compare the performance of the appropriate feature model on applying major selected subjects and; 3) develop the forecasting information system of new students to enroll in courses. The data of this study was from the student recruitment data of Chaiyaphum Rajabhat University. The researcher had used WEKA application for clustering as Simple K-Mean technique. In addition, finding association rule as Apriori algorithm and evaluate suitability by five experts, creating a model and evaluating forecasting model with the Decision Tree technique and Random Forest technique. The highest performance forecasting result of the model was developed for the information systems.

The results from the research found for all 6 clusters of the association rule it selected two interesting association rules from each cluster. It was found that the overall suitability was at the highest level. Performance model comparison was divided into 2 techniques; Decision Tree and Random Forest. As a result, the highest accuracy is Random Forest at 74.67%. It is an acceptable prediction model and applied to the development of the information systems for forecasting the number of new students and promoting proactive of public relations policies.