[3K1-OS-5a] OS-5 (1)

[3K1-OS-5a-04] Proposal of TreeMap-based Visualization Interface for Cluster Analysis of Multi-dimensional Data

〇Yasufumi Takama¹, Yuna Tanaka¹, Hiroki Shibata¹ (1. Tokyo Metropolitan University)

Keywords:clustering, visual analytics, TreeMap

This paper proposes a visualization interface for supporting cluster analysis of multi-dimensional data. It is important to grasp data distribution in a target dataset for machine learning and cluster analysis. When dealing with multi-dimensional data, dimensionality reduction algorithms are usually applied to project original data to 2-dimensional space. However, dimensionality reduction has a problem of losing characteristics of data in the original space. Instead of using dimensionality reduction algorithms, the proposed interface applies Fuzzy c-means to target data and visualizes the clustering result based on the highest membership values with TreeMap. In addition to that, membership values to other clusters are also visualized. The effectiveness of the proposed interface is discussed on the basis of the result of a user experiment.