Development of the recommendation system for the selection of major in Information and Communication Technology using decision tree techniques

Abstract

     The objectives of this research were: 1) to develop a recommendation model for choosing a major in Information and Communication Technology; 2) to compare the performance of decision tree methods such as Iterative Dichotomiser 3, C4.5, and Classification and Regression Trees (CART); and 3) to study the users’ satisfaction with the developed system. The data used to develop this model included academic results from 10 subjects: Introduction to Computer and Computer Architecture, Introduction to Programming, Statistics and Quantitative Analysis, Data Structure and Algorithms, Object-Oriented Programming, Discrete Mathematics, Database System, Web Programming, User Experience/User Interface Design, and Data Communication and Internetworking. Data imbalance was addressed using the SMOTE method. After that, the models were constructed using Iterative Dichotomiser 3, C4.5, and CART. The system was developed as a web application using PHP with a MySQL database. The evaluation results given by 10-fold cross-validation, showed that a recommendation model for choosing a major in Information and Communication Technology developed by C4.5 provided the highest level of effectiveness with an accuracy of 93.20%, precision of 93.33%, and recall of 93.32%. The user satisfaction assessment with the proposed system was collected from 30 users through questionnaires. The results indicated that the users’ satisfaction was at a high level ( \overline{x} =3.66, SD.=0.93).