Clustering Using Region Growing on Denser Regions

• Main Author: 李健銘
• Other Authors: 林振義
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/78860710369456717785

碩士 === 明新科技大學 === 電機工程研究所 === 93 === In recent years, Clustering Algorithm is considered as a powerful method in the data mining. However most exists algorithms need users to input the parameter, for instance, k-means needs users to input the clusters of number and threshold value of distance. The user does not have enough knowledge to perform the clustering algorithm with the suitable parameter values. The concept of clustering algorithm is basically made on the denser region. First of all, divide the data space into several cells of the same interval, and then calculate the size of density of the cell. Through the density on there regular areas to static the characteristic to figure out the denser zoom, this also represent different clusters. Then use the region-growing technique to find out the correct cluster of shape. This algorithm does not need users to input the parameter and the shape of cluster is also arbitrary. On top of this, it takes shorter tine to execute the algorithm.