Prediction of consumer purchase behaviour using Bayesian network: an operational improvement and new results based on RFID data

Abstract

The prediction of consumers' purchase behaviour has been extensively investigated because accurate predictions assist managers and retailers in meeting customer needs and achieving profitability. This article presents two contributions to the consumer purchase behaviour research. First, the author describes new in-store behaviour data - radio frequency identification RFID data. An RFID tag attached to a customer's shopping cart can monitor and record the in-store behaviour e.g., location coordinates and elapsed time of that customer at any time. This article refers to in-store behaviour as 'stay time' and applies it to a time-based prediction of purchase behaviour. Second, the author reveals a non-monotonic relationship between purchase behaviour and stay time. For this purpose, the author proposes an operational approach to the construction of a Bayesian network BN to predict purchase behaviour. This article experiments a new perspective on the improvement of purchase decision-making predictions in contrast with the traditional hypothesis.