A new knowledge-based constrained clustering approach: Theory and application in direct marketing |
| |
Affiliation: | 1. Department of Decision Sciences and Information Management, KU Leuven, Naamsestraat 69, B-3000 Leuven, Belgium;2. School of Management, University of Southampton, Highfield, Southampton SO17 1BJ, United Kingdom;3. Vlerick, Leuven-Gent Management School, Reep 1, B-9000 Gent, Belgium;1. Mechanical Engineering Department, South Dakota School of Mines and Technology, United States;2. Mechanical and Aerospace Engineering Department, Missouri University of Science and Technology, United States;1. Electrical and Computer Engineering (ECE), Concordia University, Montreal, QC H3G 2W1, Canada;2. Concordia Institute for Information Systems Engineering (CIISE), Concordia University, Montreal, QC H3G 2W1, Canada |
| |
Abstract: | Clustering has always been an exploratory but critical step in the knowledge discovery process. Often unsupervised, the clustering task received a huge interest when reinforced by different kinds of inputs provided by the user. This paper presents an approach giving the possibility to incorporate business knowledge in order to guide the clustering algorithm. A formalization of the fact that an intuitive a priori prioritization of the variables might exist, is presented in this paper and applied in a direct marketing context using recent data. By providing the analyst with a new approach offering different clustering perspectives, this paper proposes a straightforward way to apply constrained clustering with soft attribute-level constraints based on feature order preferences. |
| |
Keywords: | Data mining Constrained clustering Customer profiling Business knowledge Direct marketing |
本文献已被 ScienceDirect 等数据库收录! |
|