Demand Planning and Forecasting in the High Technology Industry Based on Hybrid CEEMDAN-GA-BPNN Model

Since it enables decision-makers to foresee the circumstances and scenarios surrounding a subject before acting, forecasting is essential in planning and decision-making across all fields. In order to provide precise predicted values, numerous forecasting methods have been created. Consumer price indices are useful for many things, but one of the most important is keeping everyone up-to-date on price changes that impact the Yemeni economy as a whole. Following a logical progression, we cover preprocessing, feature extraction, and training the model. Conventional methods of data preparation, such wavelet and Fourier transform technology, have a few drawbacks. Feature extraction makes use of PSO, a popular bio-inspired approach to optimization problems involving continuous and discontinuous functions. A CEEMDAN-GA-BPNN approach was used to train this model. This method appears antiquated when contrasted with GA and CEEMDAN. There is little doubt that the data demonstrates an improvement, with an impressive accuracy percentage of 96.55%.