Rough clustering of destination image data using an evolutionary algorithm |
| Presenter: |
Kevin Voges, Associate Professor, University of Canterbury, NZ |
| When: |
Friday, 29 May 2009 |
| Where: |
Room 430, Level 4, Joyce Ackroyd Boadroom, UQ St Lucia Campus |
| Time: |
10.30am - 12.00pm |
| Cluster: |
Marketing Cluster |
| Biography: |
Dr Kevin Voges is an Associate Professor in Marketing and Head of the Department of Management at the University of Canterbury, New Zealand, where he teaches market research and quantitative research methods in the MBA and Honours Programmes. Dr Voges has consulting experience in education, organizational development, business planning, and market research. He has taught research methods courses in psychology, education, business, and marketing for the past thirty years.
His research interests include the application of concepts and techniques from computational intelligence, to marketing theory and practice specifically, and to business generally. He has published in psychology, sport marketing, consumer behaviour, and computational intelligence. He has co-edited a book Business Applications and Computational Intelligence, published by the Idea Group, and has published in the Journal of Marketing Communications, the European Journal of Marketing, the Journal of Travel and Tourism Marketing, and the Journal of Advertising. |
| Workshop Paper: |
Download paper |
| Abstract: |
The technique of cluster analysis is fundamental in traditional data analysis, with many clustering methods developed, including the commonly used k-means approach. This approach is dependent on initial starting points and requires the number of clusters to be specified in advance. The rough clustering algorithm described in this presentation is able to overcome these limitations.
The concept of rough sets (also known as approximation sets), was introduced by Pawlak (1982, 1991), and is based on the assumption that with every record in the information system (the data matrix in traditional data analysis terms), there is associated a certain amount of information. This information is expressed by means of attributes (variables in traditional data analysis terms), used as descriptions of the objects. See Pawlak (1991) or Munakata (1998) for comprehensive introductions.
Rough clusters are a simple extension of the notion of rough sets. Clusters can be defined by a lower approximation (containing objects that only belong to that cluster) and an upper approximation (containing objects that belong to more than one cluster), in a similar manner to rough sets. This approach allows for multiple cluster membership for objects in the data set.
The presentation describes the template, the data structure used to describe rough clusters. It provides an overview of the evolutionary algorithm used to develop viable cluster solutions, consisting of an optimal number of templates. These cluster solutions provide easily interpreted descriptions of the clusters. This evolutionary algorithm based rough clustering algorithm was tested on a large data set of perceptions of city destination image attributes. |
Join our staff, students and alumni attending workshops presented by visiting academics on their area of research expertise.
For further information please contact our Workshop Administrator.
|
|
