Exploring Students' Attitudes Towards Robotics Technology in STEM Education: A Machine Learning-Based Prediction Analysis

Abstract

The aim of this paper is to predict the extent to which students will accept robotics technology in future teaching plans. The study involves students engaging in a two-week project-based learning program on robotics technology, after which they will decide whether it is suitable for them. To achieve this, the paper uses several prediction algorithms based on Machine Learning (ML), including Linear Regression (LR), Ridge Regression (RR), Elastic Net (EN) regression, and K-Nearest Neighbor (KNN) models. The sample group consists of 35 students from STEM high schools in Egypt. The research findings indicate that the students have gained confidence and satisfaction in robotics technology and believe it will have significant value in the future. The prediction results obtained from the ML regression models support these findings. The dataset used in the study includes seven input features and one target output representing the total prediction values of accepting and rejecting robotics technology in STEM schools. The regression models LR, RR, EN, and KNN achieve determination coefficients (R2) of 99.99%, 98.80%,99.20%, and 85.60%, respectively.