A big data and machine learning based consumer purchase decision behavior prediction system using Mc-MBGRU

Consumer purchasing behavior shows how people choose and assess and buy products. Businesses achieve their goals through accurate behavior prediction because it provides them with information about buying patterns, which helps them to improve their marketing efforts and boost customer satisfaction and drive increased revenue. Existing studies, however, lack a Dynamic Trust-aware Cross-Channel Analysis framework for behavior prediction. This research presents its solution through implementation of an approach that uses MC-MBGRU to enhance its predictive capabilities. The process starts with data collection and data preprocessing which is followed by data augmentation through SMOELITE and the identification of outliers through the Z-Score method. Attribute extraction and segmentation with DVRBSCOCAN, based on demographics and product categories, generate consumer groups. The researchers develop correlation heatmaps and Seaborn pair plots to extract features which they use to select optimal features through CKOA. These features are used to train with MC-MBGRU, achieving 99.25% accuracy in predicting consumer purchase decision behavior.