Wavelet Decomposition and Machine Learning Technique for Predicting Occurrence of Spiders in Pigeon Pea

Influence of weather variables on occurrence spiders in pigeon pea across locations seven agro-climatic zones India was studied addition to development forecast models with their comparisons performance. Considering the non-normal and nonlinear nature time series data spiders, non-parametric techniques were applied developed algorithm based combinations wavelet–regression wavelet–artificial neural network (ANN) models. Haar wavelet filter decomposed each extract actual signal from noisy data. Prediction accuracy models, viz., multiple regression, wavelet–regression, wavelet–ANN, tested using root mean square error (RMSE) absolute percentage (MAPE), indicated better performance wavelet–ANN model. Diebold Mariano (DM) test also confirmed that prediction model, hence its use conjunction values pest–defender ratios, would not only reduce insecticidal sprays, but add ecological economic value integrated pest management insects pea.