Unsupervised clustering to identify jury demographics with varying preferences

In the study of product sound quality, most approaches used to analyze the result of the jury study assume that preference within the jury population is universal. This assumption holds true in some cases, but there are often scenarios where preference varies significantly with demographic. Identifying the presence of and partitioning the jury members with these varying criteria for preference is seldom a trivial exercise. For this reason, a method that can be used to evaluate a jury population and identify subgroups based on the juror voting patterns would be a useful tool. To address this problem, this paper presents an approach to infer the number of subgroups and to classify the jury members into the appropriate subgroups. The method is demonstrated in this paper using both a K-means clustering and a Ward's clustering algorithm. In addition to providing a tool to improve the quality of the jury preference models, jury subgroups also have implications on brand imaging for target audiences. Identifying subgroups of consumers with differing preferences provides marketing groups with additional knowledge that can be used for targeted brand imaging.