An Improved Mathematical Model to Predict and Evaluate the Buying Assessment of Customers During a Pandemic Situation in Kerala

Abstract

Nowadays, the knowledge-based economy has attracted more attention, especially online purchasing platforms where all activities and customer reviews are recorded. Implicit information from the records might be extracted using machine learning techniques. Industries and companies use the information to effectively comprehend consumer behaviour, possibilities, and threats. The coronavirus (COVID-19) outbreak has significantly impacted many facets of our daily lives, including shopping habits. Managers in the retail, distribution chain and public sector could benefit from determining customer behaviour regarding electronic devices. The research had seen online purchasing before the coronavirus epidemic, but its prevalence surged significantly during the illness. Due to COVID-19's high communication, The researchers must be aware of social isolation and personal well-being difficulties. These problems directly affect how consumers behave when shopping online. A Consumer Buying Assessment during Pandemic (CBAP) is proposed in this research. A proposed methodology is put out in this research using ml algorithms to forecast customer behaviour. Five individual classifications and combinations with bagging and enhancing are investigated on a dataset from an online retailer. According to the findings, the choice tree ensemble method with Bagging produced a model that was 95.3% accurate in predicting customer behaviour. Additionally, correlation analysis is carried out to identify the key elements affecting the number of online purchases made during the coronavirus epidemic.