Comparative analysis of wheat yield prediction through artificial intelligence, simulation modelling and statistical analysis in Central Punjab

Recognizing the paramount importance of wheat in agricultural landscapes, crop yield forecasting assumes a pivotal role in various aspects of farming planning and management.Beyond its implications for domestic food supply, accurate yield forecasts contribute to informed decision-making in international food trade, facilitate ecosystem sustainability efforts, and play a crucial role in the broader context of agricultural practices (Pandey and Sinha, 2006).Integration of artificial intelligence including ARIMA, LASSO, ANN, ELNET, MLP, ELM (Gomez et al., 2021) and simulation modelling like CERES-wheat model (Ritchie et al., 1988), reveals a high level of accuracy in predicting wheat yield, providing valuable insights into the relationships between crop yield and weather parameters.These simulation models enhance predictive precision, facilitating informed decision-making and sustainable agricultural practices.Researchers have enhanced the precision of predicting wheat yield in Ludhiana, by developing different models (Singh et al., 2021).The daily weather data of different parameters (maximum temperature, minimum temperature, morning relative humidity, evening relative humidity, rainfall and bright sunshine hour) for the years 2001 to 2022 were collected from the Agrometeorological Observatory of Department of Climate Change & Agricultural Meteorology, Punjab Agricultural University, Ludhiana.Correspondingly, wheat yield data for Ludhiana district was collected from statistical abstracts of Punjab for the years 2001-2022.The weekly weather variables were generated using daily data by averaging the daily maximum temperature (Tmax) and daily minimum temperature (Tmin), averaging the morning relative humidity (RH1) and evening relative humidity (RH2), and summing up the rainfall (RF).These weekly aggregated variables were utilized for subsequent analysis and for generating the weather indices following the method described by Ghosh et al., (2014).