A Dissimilarity Measurement Method for Hierarchy Variable with Different Structures
A hierarchy structure of variable is a regular form to arrange qualitative attributes, a tree with certain properties. In clustering analysis it is a difficult work to calculate the dissimilarity degree between any two qualitative attributes in a hierarchy tree. Some dissimilarity metric methods have been proposed to solve this problem but they don’t involve the same qualitative attributes with different hierarchy structures. A dissimilarity metric method for this type is proposed in this paper which can reflect the influence of different structures. Moreover, dissimilarity metric for the hybrid variables including several traditional types and one hierarchy type is designed, and a clustering algorithm based on this metric is implemented.
KeywordsDissimilarity metric Hierarchy variable Hierarchy structure Clustering analysis
Authors gratefully acknowledge the editor and anonymous reviewers for their valuable comments and constructive suggestions. This research is supported by the Soft Science Foundation of Shandong Province, China (Grant No. 2011RKGA2003).
- 1.Han J, Kamber M, Pei J (2011) Data mining: concepts and techniques, 3rd edn. Morgan Kaufmann, Waltham, pp 65–76Google Scholar
- 2.Pei-Ying Y, Da-Lin W (2007) Dissimilar metric and clustering for hierarchy variable. Comput Sci 34(10):214–216Google Scholar
- 3.Joshi R, Patidar A, Mishra S (2001) Scaling k-medoid algorithm for clustering large categorical dataset and its performance analysis. In: International conference on electronic computer technology, Kanyakumari, India, pp 117–121Google Scholar
- 4.Serguei Levachkine, Adolfo Guzmán-arenas (2004) Hierarchies measuring qualitative variables. Lecture notes in computer science 2945. Springer, Seoul, South Korea, pp 262–274Google Scholar