Predicting Crops Based on Soil and Environmental Characteristics Using ARIMA and LSTM Model

Food producers are under pressure to meet rising demand as the world's population and natural resources diminish. Overuse of pesticides and fertilizers has caused soil erosion and land degradation. Future food needs will require environmentally sustainable farming strategies. The suggested strategy includes model training, feature extraction, sensor integration, and model preparation. Data preparation involves processing data before storing it machine-readable. SFFS and RFE select features. The model was trained using a hybrid ARIMA-LSTM technique, which blends deep learning and time-series forecasting. The ARIMA-LSTM model outperforms solo ARIMA and LSTM models with an average accuracy of 93.54%. Increased accuracy shows the model's capacity to anticipate sustainable agriculture approaches. Data-driven and environmentally sustainable ways are needed to overcome global food production concerns, according to this study. When paired with good preprocessing and feature selection, ARIMA-LSTM can boost agricultural productivity while minimizing environmental impact.