Prediction of Paddy Yield based on IoT Data using GRU Model in Lowland Coastal Regions

Agriculture is the primary driver of economic growth in India, one of the world’s most populous countiies. Nearly sixty percent of crop productivity relies on soil fertility and environmental conditions. Therefore, regular soil tests and identifying weather conditions are essential to ensure crops thrive. Due to a lack of knowledge and awareness of the soil, many errors are made in the fertilizing and planting processes. The chemical, physical, and biological qualities of soil decide whether it will produce fertile crops. The nutritional content of the earth, as well as its pH level, can be determined using soil testing. Previously most of the research in yield prediction based on different parameters like environmental and type of soil is taken place only once before planting. Instead, soil characters and environmental parameters should be analyzed regulaily and prompt actions should be taken. In this paper, IoT based framework is used to measure the soil parameters like nitrogen(N), Phosphorus (P), and Potassium (K$\rangle$. along with soli parameters, collected the environment variables and made an analysis using a Deep learning-based modified GRU model to predict the right combinations of the sola and environmental parameters. The proposed model performs better in predicting paddy yieldthan the state of art models.