INTELLIGENT FORECASTING OF AGRICULTURAL PRODUCTION THROUGH ADVANCED DATA MINING ALGORITHMS

Abstract

Precise estimation of farm production is a crucial component of agricultural planning, food safety, and financial stability, particularly in an agrarian economy like Pakistan's. This paper examines the practice of data mining and machine learning to predict crop yields in different environmental and agronomic conditions. Several predictive models, comparable linear regression, multinomial naive Bayes, decision tree, XGBoost Regressor, Stochastic Gradient Descent (SGD) Regressor, kernel ridge regression, Elastic Net, Bayesian ridge regression, Gradient Boosting Regressor, and Support Vector Regression (SVR), were developed and tested on historical agricultural data that included type of crop, cultivated land, average temperature, rainfall, and pesticide use. Accuracy, mean absolute error (MAE), mean squared error (MSE), and root mean squared error (RMSE) were also measures of model performance. According to the experimental results, the Decision Tree model is the best in the comparison with the other methods, with an accuracy of 93.27, an MAE of 6.73, and an RMSE of 2.59. The outcomes suggest that decision tree-based techniques are very successful in the prediction of crop yields and could be of use in giving decision support to the farmers and policymakers. A decision tree is an algorithm of managed machine learning, which applies a hierarchical, tree-based framework of decision-making and prediction.