Enhanced GRU-BiLSTM Technique for Crop Yield Prediction

Agriculture is the major source of food and significantly contributes to Indian employment, and the economy is intricately tied to the outcomes of crop management, where the final yield and market prices play crucial roles. The final yield and the market price completely determined the outcome of crop management or agriculture in India. Real-time observation emerges as a critical determinant of overall crop production success. Recognizing the significance of insightful analysis and precise crop yield predictions for effective farming practices, this study proposes an enhanced model to address the imperative of accurate yield forecasting. The pre-processing steps of the proposed model include Min-Max normalization, deletion of irrelevant data, and addition of missing values. The pre-processed data is then subjected to feature extraction using an Improved Shearlet transform (IST). After feature extraction, feature selection is done using an Enhanced multi-objective Grey Wolf optimization (EMGWO) technique. Finally, the prediction is made using an enhanced Gate Recurrent Unit with a Bidirectional LSTM (GRU-BiLSTM) model. This enhanced the accuracy (97