Volume 15, No 1, 2018
Developing A Prediction Model For Academic Success Of Physics Major Students: An Analysis Employing Data Mining Technique
Vinchail A. Siason
Abstract
In this paper, factors that influenced the success of Physics students in a Philippine university was determined. Specifically, the study aimed to find out whether significant correlation exists between the 1st Year grades of the students in Mathematics, Science, English and College Entrance Test (CET) rating with their retention in the BS Physics program. It also generated a predictive model of success for the students and determined which variable is the greatest predictor for the student study outcome. The study employed the descriptive correlation method research design. Data was analyzed using Point-biserial Correlation and Logistic Regression analysis. Results using PointBiserial Correlation at p<0.05 disclosed that all of the cognitive variables (First year subjects and CET Results) were found to be correlated with the student study outcome at varied measures. Binary Logistic Regression analyses generated the prediction model and deduced that the grades in Calculus with Analytic Geometry and English have the most significance in predicting the success of the Students in the BS Physics program in the specific university while the CET Results and Algebra were found to have no influence in predicting the study outcomes of the students. The generated statistical model is found to be valid and acceptable with overall percentage accuracy of 84.8%.
Pages: 132-147
Keywords: Student Retention, Data Mining, Logistic Regression, Physics students