Evaluating Machine Learning Models for Pan-Indian Agricultural Yield Prediction

Tambe S.L.^1*, Dr. Dashore Pankaj²

---

ABSTRACT

This paper presents a comparative evaluation of the classical machine learning models and the deep machine learning models in the prediction of the crop yield using a pan-Indian agricultural dataset. A government-based composite dataset comprising 19,690 samples was built and a strict feature engineering pipeline was adopted to deal with quality of data and multicollinearity. There were four models namely Linear Regression (LR), Decision Tree (DT), Artificial Neural Network (ANN), and a Convolutional Neural Network – Long Short Term Memory (CNN-LSTM) hybrid that were systematically tested. However, findings showed that the Linear Regression simple model gave a better predictive power with a score of R² of 0.9942. It concludes the paper by finding that in the case of structured tabular data at national scale computationally efficient classical models combined with careful feature engineering can be more effective at crop yield forecasting than more complicated deep learning methods.

Keywords: Machine Learning (ML), Linear regression (LR), Decision Tree (DT), Artificial Neural Network (ANN), Convolutional Neural Network – Long Short Term Memory (CNN-LSTM)

How to cite this article: Tambe SL, Dashore P. Evaluating Machine Learning Models for Pan-Indian Agricultural Yield Prediction. Int J Drug Deliv Technol. 2026;16(3): 747. DOI: 10.25258/ijddt.16.3.82

Source of support: Nil.

Conflict of interest: None