Consumer Behavior in the Post-Pandemic Era: Insights from Machine Learning and AI-Driven Analytics

COVID-19 has been the disruptor in the world, and the society has been changed from one perspective of the consumer behavior in terms of digital platforms and their preferences for safety, convenience and sustainability. It is therefore very important for any business to acquaint itself with these changes so that may be in a position to effectively compete for the market within the current ever-changing business environment. Conventional consumer behavior models, on the other hand, rely on fragmented data sources and analytical tools that may not reflect the current post-pandemic consumption patterns. Such models are also incapable of incorporating multiple data types and can also fail to address changing social factors. In response, a Multi-Modal Consumer Behavior Prediction Framework (MCBP) that considers a vast array of data sources is introduced. BERT and GNNs are used to give a detailed, real-time interpretation of consumer behaviour accomplished by the MCBP framework. It uses BERT to extract and analyze sentiment and topics of the textual data and GNNs to capture influence and peer effects on purchase decisions. The findings of the study show the effectiveness of the proposed framework with an accuracy of 99%. 5% and the implementation is done using Python. The high level of accuracy shown assures the success of the MCBP model in measuring and forecasting consumers' behaviors in the post-pandemic world.