Project: Individual Consumer Modeling

Loyalty card programs at many grocery chains have resulted in the capture of millions of transactions and purchases directly associated with the customers making them. Traditionally, most of the data mining work using retail transaction data has focused on approaches that use clustering or segmentation strategies. This is usually done to overcome the data sparseness problem and results in systems that are able to overcome the variance in the shopping behaviors of individual customers, while losing precision on any one customer. We believe that given the massive amounts of data being captured, and the relative high shopping frequency of a grocery store customer, we can develop individual consumer models that are based on only a single customer's historical data. Our hypothesis is that by utilizing the detailed transaction records to build separate classifiers for every unique customer, we can improve on the performance of clustering and segmentation approaches.

We formally frame the shopping list prediction as a classification problem. This results in a system that has to build billions of models (several models for each customer, product pair). Our results show that we can predict a shopper's shopping list with high levels of accuracy, precision, and recall. We believe that this work impacts both the data mining and the retail business community. The formulation of shopping list prediction as a machine learning problem results in algorithms that should be useful beyond retail shopping list prediction. For retailers, the result is not only a practical system that increases revenues by up to 11%, but also enhances customer experience and loyalty by giving them the tools to individually interact with customers and anticipate their needs.

People:

Papers:

Learning Individual Consumer Models for Personalized Promotions: A Data Mining Case Study.
Chad Cumby, Andrew Fano, Rayid Ghani, and Marko Krema.
Workshop on Data Mining for Business — held with the European Conference on Machine Learning (ECML/PKDD 2005).

Predicting Customer Shopping Lists from Point-of-sale Purchase Data
Chad Cumby, Andy Fano, Rayid Ghani and Marko Krema
10th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
August 2004
Seattle, Washington

Building Intelligent Shopping Assistant Using Individual Consumer Models
C. Cumby, A. Fano, R. Ghani and M. Krema
Proceedings of the 2005 International Conference on Intelligent User Interfaces
January 9-12, 2005
San Diego, California