A1057
Title: Tree-structured clustering model using projection pursuit method and their explanation
Authors: Eun-Kyung Lee - Ewha Womans University (Korea, South) [presenting]
Abstract: The purpose is to propose a tree-structured clustering model using Projection Pursuit. Projection Pursuit is a dimension reduction methodology that is often used for exploratory analysis of high-dimensional data. Finding the optimal projection vector containing the pattern by the user-defined projection pursuit index is possible. The projection pursuit indices are investigated for unsupervised learning, combined these indices with the tree-structured model for projection pursuit, and proposed the projection pursuit clustering tree. Ten data sets with known class information were used to check the model's performance, and the results for each projection pursuit Index were compared.