Forecasting E-Commerce Sales Adoption Based on DE-ELM-RGSO Approach

Most agricultural supplies companies have established corporate Web presences and Internet strategies by 1999. Though some purchases were made directly by customers, the great majority were made by vendors (channel partners) and not end users. Business owners that witnessed logistical and stock management improvements through e-commerce were more likely to adopt the practice. It was hoped that farmers would take to e-commerce because of the streamlined experience it provided them with researching and buying goods. The proposed methodology begins with the preliminary processes of preprocessing, feature extraction, and model training. After carefully evaluating the raw data, four stages of preprocessing were implemented: treatment of missing values, rectification of skewed distribution, data analysis of outliers, and application of normalization. Both linguistic and informational features are employed in feature extraction. In order to train DE-ELM-RGSO models, feature extraction must first take place. When compared to DE and RGSO, the two most popular alternatives, the proposed technique emerges as the clear winner. By implementing the plan, they increased accuracy to 95.76 percent.