Cap-DiBiL: An Automated Model for Crop Water Requirement Prediction and Suitable Crop Recommendation in Agriculture

Abstract In this technological era, several approaches used to provide the information about suitable crop recommendation means which is for soil. Some of depends on IoT smart agricultural-devices gather from surrounding area. However, collection data are predict crops details but it not efficient better performance. Therefore, proposed model uses various techniques improve performance efficiently. steps involved in as pre-processing, feature extraction, selection, water requirement prediction and recommendation. Initially, collected dataset pre-processed using normalization, missing value imputation one-hot encoding. Then, extract features Gated Residual autoencoder (GRA) model, whereas optimal selected Chaotic Northern Goshawk Optimization (ChaNgo) algorithm. Based farmland details, due market price carried out a novel hybrid deep learning called Channel capsule-assisted stacked dilated Bi-LSTM (Cap-DiBiL). The channel capsule network predicts recommendations such millets, rice other crops. Then analyses compares with existing prove model’s enhancement. improved accuracy 98.18% predicting Precision, Recall F1 score also enhanced 98.31%, 98.20%.