A Shrinking-Based Approach for Multi-Dimensional Data Analysis Yong Shi, Yuqing Song, and Aidong Zhang   View Paper (PDF)   Return to Data Mining (Session B2) Abstract in the real world. This data reorganization concept can be applied in many fields such as pattern recognition, data clustering and signal processing. Then, as an im- portant application of the data shrinking prepro- cessing, we propose a shrinking-based approach for multi-dimensional data analysis which con- sists of three steps: data shrinking, cluster de- tection, and cluster evaluation and selection. The process of data shrinking moves data points along the direction of the density gradient, thus gener- ating condensed, widely-separated clusters. Fol- lowing data shrinking, clusters are detected by finding the connected components of dense cells. The data-shrinking and cluster-detection steps are conducted on a sequence of grids with different cell sizes. The clusters detected at these scales are compared by a cluster-wise evaluation measure- ment, and the best clusters are selected as the final result. The experimental results show that this ap- proach can effectively and efficiently detect clus- ters in both low- and high-dimensional spaces. ©2004 Association for Computing Machinery