Optimizing Demand Forecast for E-commerce Sales Platforms Based on RNN-RBM Methodology

Due to technical advancements in this digital era, particularly in the e-commerce sector, the corporate paradigm has changed substantially. A common platform for advertising products is online marketplaces, sometimes known as ecommerce platforms. Businesses in the increasingly competitive e-commerce space require a strong marketing plan to boost sales conversion and differentiate between the competitions. Selected features, preprocessing, and training of models all depend on correct sequencing. The proposed approach utilized the SG smoothing filter during the preprocessing phase. Principal component analysis (PCA) is a statistical method that can be employed in feature selection to decrease the dimensionality of a dataset that contains numerous associated variables. Precise control over the qualities is necessary for RNN-RBM training. This approach appears to be far more cutting-edge than the current RBM and RNN algorithms. A significant improvement in accuracy was noted in the results, which reached 96.33%.