An Innovative Method for Paddy Yield Prediction Based on DCNN-ELM Approach

The majority of a country's population relies on agriculture for their livelihood. Water shortage, uncertainty in market prices due to supply and demand, erratic weather, and inaccurate crop predictions are just some of the new realities that farmers must adjust to. Environmental conditions, management tactics, crop genotype, and yield all interact in complex ways, making crop yield prediction difficult. This is especially true for paddy cultivation. Researchers utilize to forecasting the paddy yield through statistical methods, yet they failed to obtain greater precision due to several causes. Within the suggested methods, preprocessing, feature selection, and model training are all included. They employ adaptive data cleaning and denoising auto encoder in data preprocessing. All of the missing values in the input dataset are filled in using the pre-processing technique. They employ feature selection the CFS procedure uses statistical measurements as evaluation criteria to determine the quality of feature subsets. DCNN-ELM trains the models after the attributes are gathered. Compared to the DCNN and ELM techniques, the suggested strategy yields significantly better results. At 96.33 percent, this technique has an incredibly high probability of success.