A Comparative Study of Bootstrap Application to Subjective Clustering

Abstract

In order to develop successful products in today's competitive market, it is essential to analyze customer needs accurately. Since the number of customer needs collected using marketing survey tends to be large, it is necessary to group similar needs and identify a small number of representative needs. Subjective Clustering is a statistical method that uses Hierarchical Clustering algorithm to group similar customer needs into clusters. This paper compares Subjective Clustering (SC) and Bootstrap applied to Subjective Clustering (BS-SC) for determining how accurately these methods estimate population primary clusters. in order to investigate inherent characteristics in the two methods, this paper proposes a simulation-Based comparative study. the result suggests that BS-SC estimates population primary cluster more accurately than SC when the sample size is small relative to population size; however, as sample size increases, SC estimate is more accurate than BS-SC. Copyright © 2006 by ASME.

Department(s)

Mechanical and Aerospace Engineering

Keywords and Phrases

Bootstrap; Comparative study; Customer needs; Hierarchical clustering; Subjective clustering

International Standard Book Number (ISBN)

978-079183784-9

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2024 American Society of Mechanical Engineers, All rights reserved.

Publication Date

01 Jan 2006

Share

 
COinS