Using SOM to Mine Product Features from Free-Text Customer Reviews

Abstract

This study examines how the Self-Organizing Map (SOM) technique can be used to identify product features from free-text customer reviews. A novel SOM display named “Attribute Accumulative Matrix” is presented. To verify the validity of the proposed approach, 22157 restaurant reviews are collected and product features of catering industry are identified. The experiment results show that this approach can effectively identify the product features.