Prediction Model for Precision Agriculture Using Machine Learning

Precision farming has drawn a lot of interest because it provides farmers with technology that enables them to make the best decisions possible using accurate data from IoT-enabled sensors, actuators, robot drones, satellite photos, meteorological data, and other real-time data feeds. By using artificial intelligence and machine learning to the gathered data, governments and farming agencies can make better decisions on important aspects of agriculture, such as soil characteristics, weather, and crop yield forecast. To this end, this work study various supervised learning models such as Linear Regression (LR), Logistic Regression (LGR), Decision Tree (DT), Support Vector Machines (SVM), Support Vector Regression (SVR), Naive Bayes (NB), Stochastic Gradient Descendent (SGD), and Random Forest (RF) for classification and regression on the agriculture datasets downloaded from Kaggle website. The performance of the classification algorithms evaluated using metrics such as accuracy, precision, recall, and F1-score, where as the regression algorithms are evaluated using metrics such as R2 score, and Root Mean Square Error (RMSE) for selecting the best model. In comparison to other supervised learning algorithms, it has been found that RF has the best accuracy (99%) and the lowest RMSE. A web application built with the Python Flask framework and integrated with the recommendation system allows users to interact with the RF model for decision support.