Impacts of Learning Orientation on the Modeling of Programming Using Feature Selection and XGBOOST: A Gender-Focused Analysis

Abstract

In programming, modeling is a generalized explanatory system that organizes key characteristics of a specific matter or object of interest using computer science concepts. Modeling is integral for both automation design in programming education and communication in the collaborative process. This study investigates the effect of learning orientation on the modeling stage based on gender. The study includes 756 male and 688 female elementary-school students. We analyzed the results of XGBOOST by extracting the influential characteristics from feature selection along with the basic statistics. As a result of the study, it was confirmed that learners, regardless of gender, had the largest gap in modeling and that this was the stage at which differences occurred in programming education. For male students, the process of collecting data for modeling or devising a solution was found to be an important learning method. This shows that it is necessary to create an environment to focus on activities that derive solutions from the collected data along with strengthening information retrieval education. Although female students showed a similar tendency to male students, the process of cooperating with friends as a differentiating factor was found to be an important learning method. It seems necessary to apply teaching and learning methods that can strengthen team projects that can collaborate with friends. The findings could serve as a reference for teaching and learning design and operation for effective programming education.

Keywords

• learning orientation
• programming
• gender
• feature selection
• XGBOOST