ABSTRACT
The research paper explores the application of machine learning models for agriculture usage in the cases of the LSTM, KNN and Extreme gradient boosting XGBoost. We also develop machine learning models based on historical data of weather station to examine the ability of those models to capture the critical weather patterns that are relevant for agriculture planning and management. The dataset is modified through pre-processing and feature engineering to use for training and evaluating of the machine learning models. To make the comparison, the performance results such as accuracy, MAE and RMSE and the computational efficiency, quality, and performance of machine learning models. The results reveal the differences in the variables to be counted and the selected models are unable to give accurate predictions it was noted. To summarize, the machine learning used. In the agriculture sector has many advantages and strength to address decision making and exemplifies resource allocation, management strategies, yield, and mitigation optimization. In the Research on LSTM, KNN and XGBoost. LSTM is best preferred in predicting and forecasting weather pattern having least and best (MAE and RMSE) value 0.02,0.045 and XGB 13.1 and 4.2. K-nearest neighbours have an accuracy of over 75% and acceptable.
