admin@publications.scrs.in   
Computational Intelligence and Machine Learning

Enhancing STEM Education through Data Science: Predicting Electric Motor Performance with Linear and OLS Regression

Authors: Sundeep H Deulkar and Sagarika Pramod


Publishing Date: 30-04-2025

ISBN: 978-81-975670-5-6

DOI: https://doi.org/10.56155/978-81-975670-5-6-9

Abstract

This study bridges practical engineering tasks with data science analysis to enhance STEM education for high school students. In a hands-on exercise involving 60 teams of 12thgrade students, participants constructed electric motors to achieve maximum rotations per minute (RPM) using various values of magnetic field strength (B), current (I), number of coil turns (N), and coil cross-sectional area (A). We utilized this data to build and train linear regression and Ordinary Least Squares (OLS) regression models to predict RPM and power based on the measured values of B, I, N, and A. The analysis involved cleaning the data, performing regression, and evaluating the model’s performance using metrics such as R² values, p-values, Mean Squared Error (MSE), and model comparison criteria including AIC and BIC. The linear regression model for RPM yielded an R² value of 0.209 and a Mean Squared Error of 1043.14, indicating a weak fit. The OLS regression model for RPM also showed a low R² value of 0.205, with an adjusted R² of 0.046, and a non-significant F-statistic (p-value = 0.309). In contrast, the linear regression model for power resulted in an R² value of 0.450 and an MSE of 1090.11. The OLS regression model for power demonstrated a stronger fit with an R² value of 0.573, an adjusted R² of 0.487, and a significant F-statistic (p-value = 0.00136), along with lower AIC and BIC values compared to the RPM model. This approach not only facilitated practical learning but also demonstrated the application of data science in analysing and optimizing engineering designs. The results highlight the critical factors influencing electric motor performance and underscore the educational value of integrating data- driven analysis into STEM projects. This study provides valuable insights for future curriculum development, emphasizing the role of data science in enhancing experiential learning.

Keywords

Electric Motor, Experiential Learning, Linear Regression, Predic- tive Modeling.

Cite as

Sundeep H Deulkar and Sagarika Pramod, "Enhancing STEM Education through Data Science: Predicting Electric Motor Performance with Linear and OLS Regression", In: Sandeep Kumar and Kavita Sharma (eds), Computational Intelligence and Machine Learning, SCRS, India, 2025, pp. 111-120. https://doi.org/10.56155/978-81-975670-5-6-9

Recent