Jinsan Yang and Byoung-Tak Zhang,
Customer Data Mining and Visualization by Generative
Topographic Mapping Methods,
In Simeon J. Simoff, Monique Noirhomme-Fraiture and Michael H. Böhlen eds., Proceedings of the International Workshop on Visual Data
Mining (VDM@ECML/PKDD2001),
Freiburg, Germany, pp. 55-66, 4 September 2001.Understanding various characteristics of potential customers is important in the web business with respect to economy and efficiency. When analyzing a large data set on the customers, the structure of data can be highly complicated due to the correlations and redundancy of the observed data. In that case a meaningful insight of data can be discovered by applying a latent variable model to the observed data. Generative topographic mapping (GTM) [2] is a latent graphical model which can simplify the data structure by projecting a high dimensional data onto a lower dimensional space of intrinsic features. When the latent space is a plane, we can visualize the data set in the latent plane. We applied GTM methods in analyzing the web customer data and compared their relative merits on the clustering and visualization with other known method like self-organizing map (SOM) [10] or principle component projections (PCA). When applied to a KDD data set, GTM demonstrated improved visualizations due to its probabilistic and nonlinear mapping.