Crop Yield Prediction Using Machine Learning Algorithms

Crop yield prediction is essential in modern agriculture for food security and economic growth. As long as accurate predictions are made by governments can prepare for food shortages, ensure that there is plenty of food to go around for the population and assist farmers in decision-making. These predictions guide better practices and seed selection. The present study aims to predict rice yield across Tamil Nadu, focusing on adapting to climatic changes to boost accuracy. This crop yield forecast enhances the livelihood of the farmers. The parameters used in the prediction are rainfall, humidity, temperature, soil profile, pesticides, and past yield data. We have made use of the public data, namely data.gov.in as well as NASA COVER, for meteorological and historical yield information. To attain accuracy, we use both satellite data and meteorological data. Preprocessing occurs before using data in predictive models. Data cleaning removes inconsistencies from data, whereas Exploratory Data Analysis helps understand the characteristics of data. We are going to utilize two advanced models of machine learning: Random Forest and Gated Recurrent Unit network. Both will be used in a proper ensemble to create the best model for prediction that yields robust performance. It manages sequential data and retains previous inputs to discover patterns. While random forest captures static relationships, GRU focuses on temporal dependencies. The combination reduces overfitting and enhances performance.