A Comparative Analysis of Regression Models for Crop Yield Prediction Based on Rainfall Data: Experimental Study and Future Perspective

The ability to anticipate crop yield is essential for agricultural planning and decision-making. Crop output is significantly influenced by rainfall, a significant environmental component. In this study, we suggest using rainfall data to forecast agricultural yield using a machine learning regression-based method. In order to train and assess several regression models, such as linear regression, decision tree regression, and support vector regression, the research makes use of historical rainfall and crop yield data from numerous agricultural locations. The data used for this study spans multiple years and contains details on agricultural yields for various crops as well as rainfall patterns. The methodology entails handling missing values and outliers in the data through preprocessing, extracting pertinent features from the rainfall data through feature engineering, and training the model through supervised machine learning. Different assessment metrics, such as mean squared error (MSE), root mean squared error(RMSE), and coefficient of determination (R-squared), are used to assess each regression model's performance. The experimental results show that machine learning regression models, with random forest regression outperforming the other models, can accurately forecast agricultural yield based on rainfall data. The suggested models can offer farmers and decision-makers useful information to help them decide on crop management practices, like fertilizer use and irrigation scheduling, to maximize crop production. In order to further increase the precision of crop production prediction models, future research can investigate the integration of other environmental and socioeconomic parameters.