Giuseppe Ieva’s research while affiliated with Ospedale Bassini - Cinisello Balsamo and other places

Retail companies are greatly interested in performing continuous monitoring of purchase traces of customers, to identify weak customers and take the necessary actions to improve customer satisfaction and ensure their revenues remain unaffected. In this paper, we formulate the customer churn prediction problem as a Predictive Process Monitoring (PPM) problem to be addressed under possible dynamic conditions of evolving retail data environments. To this aim, we propose TSUNAMI as a PPM approach to monitor the customer loyalty in the retail sector. It processes online the sale receipt stream produced by customers of a retail business company and learns a deep neural model to early detect possible purchase customer traces that will outcome in future churners. In addition, the proposed approach integrates a mechanism to detect concept drifts in customer purchase traces and adapts the deep neural model to concept drifts. Finally, to make decisions of customer purchase monitoring explainable to potential stakeholders, we analyse Shapley values of decisions, to explain which characteristics of the customer purchase traces are the most relevant for disentangling churners from non-churners and how these characteristics have possibly changed over time. Experiments with two benchmark retail data sets explore the effectiveness of the proposed approach.

Customer loyalty is a crucial factor for retail business success. This paper illustrates an AI approach, named CENTAURO, to learn customer loyalty prediction models that may help retailers to run powerful loyalty programs and take better decisions. In particular, the proposed approach learns a classification model from the Recency, Frequency and Monetary (RFM) value of historical customer shopping data. For this purpose, the RFM model is extended to monitor Recency, Frequency and Monetary both over time and over the various categories of products purchased. Experiments performed with a benchmark dataset explore the performance of the extended RFM model in combination with several classification algorithms (e.g., Logistic Regression, Multi-Layer Perceptron, Random Forest, Decision Tree and XGBoost). Finally, we use an eXplainable Artificial Intelligence (XAI) technique – SHAP – to explore the effect of RFM values on the customer loyalty profile learned through the classification model.