UC Santa Cruz

Categorizing online consumers into Hedonic or Utilitarian (H/U) has long been known helpful for improving shopping experience and this problem has been investigated by scholars from economics and marketing domains over the last decade. Existing work focuses on identifying characteristics that best represent H/U at a qualitative level and there is a dearth of effort in the study of computational approach for the categorization problem. We are motivated to solve the categorization problem by utilizing methods developed in data science field. The present paper employs a machine learning based approach which categorizes online consumers by leveraging available consumer behavioral data. We evaluated the proposed approach on a real world e-commerce data set. The experimental result demonstrates the feasibility of classifying consumers just based on small number of high leveled behavioral data. In the future, better classification accuracy could be achieved by incorporating more sophisticated information about con- sumer.