Dominant Criteria and Its Factor Affecting Student Achievement Based on Rough-Regression Model

The ordinary least square model has been widely considered to estimate the significant factors which influence the student achievement. Some factor is qualitative type and measured using criteria or categories. However, the decisive criteria for each factor which affect to the cumulative grade point average of student cannot be determined by this model. In this paper, we are interested to build a new procedure using rough-regression model in determining the dominant criteria from each factor based on generalization of dependency attribute. Based on result, the proposed procedure is capable to investigate the dominant criteria and factors affecting student achievement, such as, language spoken with dominant criteria is “many-many”, FB friend with dominant criteria is “many” and fast food with dominant criteria is “never”. This proposed procedure is very appropriate to implement for handling categorical data.