Comparative Performance of Support Vector Machine and Random Forest Algorithms for Mango Yield Prediction in Dharwad and Kolar Districts of Karnataka

Keerthi P, Vasantha Kumari J, Lalitha V. M.

Abstract: Aims: The study aimed to develop and compare the performance of two machine learning models-Support Vector Machine (SVM) regression and Random Forest (RF)-for predicting mango (Mangifera indica L.) yields in Dharwad and Kolar districts of Karnataka, representing distinct agro-climatic zones. It further assessed the predictive capability of these models under varying climatic conditions. Study Design: A retrospective analytical study was conducted using machine learning-based regression modelling. Place and Duration of Study: The study was carried out in the Department of Agricultural Statistics, University of Agricultural Sciences, Dharwad, using secondary data on mango yield and weather parameters spanning 44 years (1980-2023). Methodology: A dataset comprising mango yield statistics and meteorological variables-rainfall, maximum and minimum temperatures, and wind speed was used. Both SVM and RF models were trained on 80% of the data and tested on the remaining 20%. Model performance was evaluated using the coefficient of determination (R?), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). Scatter plots were utilized to visualize relationships between actual and predicted yields. Results: The Random Forest model exhibited superior predictive accuracy compared to SVM in both districts. In Dharwad, RF achieved an R? of 0.513 versus 0.433 for SVM, while in Kolar, RF attained 0.760 compared to 0.079 for SVM. Scatter plots indicated that RF predictions aligned more closely with observed yields, particularly in Kolar. Conclusion: Ensemble-based models such as Random Forest outperform kernel-based SVM for mango yield prediction. Integrating long-term meteorological data with machine learning techniques enhances yield forecasting accuracy and supports climate-resilient agricultural planning.

Keywords: Mango yield, Machine learning, Support Vector Machine, Random Forest, Weather parameters

How to Cite?: Keerthi P, Vasantha Kumari J, Lalitha V. M. , "Comparative Performance of Support Vector Machine and Random Forest Algorithms for Mango Yield Prediction in Dharwad and Kolar Districts of Karnataka", Volume 14 Issue 10, October 2025, International Journal of Science and Research (IJSR), Pages: 1376-1379, https://www.ijsr.net/getabstract.php?paperid=SR251026164856, DOI: https://dx.doi.org/10.21275/SR251026164856