Job congruence classification model using decision tree induction algorithm

Abstract

Employability is defined in various ways, one of those is job congruence. The West Visayas State University Learning Assessment Center conducts Terminal Competencies Assessment to determine workplace readiness. With the use of the TCA data, this study aimed to develop a classification model that successfully classifies the level of job congruence of college graduates. Since TCA data is numerical, discretization was performed to yield proficiency levels. Missing data was addressed by using Multiple Linear Regression Analysis using the existing variables. Several Decision Tree algorithms were administered to develop the model. Performance analysis of these algorithms was evaluated. When General Competency and Specialized Knowledge scores are available, the proposed classification model has an accuracy rate of 89.158, which was generated thru SimpleCART algorithm. When the available score is only from the General Competency exam, the classification model with the highest accuracy rate was generated by the J48Graft algorithm at 89.13%. The statistical test shows that these two do not have significant difference thus, both can be utilized. There were two significant variables contributing to the level of job congruence; Technological Facility and Critical Thinking.