本論文根據均數平移的想法，針對LR-型模糊數提出一穩健的聚類分析演算法。此演算法不需事先給定分群數且不受起始值影響。經由數值模擬顯示：所提之演算法能有效地將資料分群且分群結果不受極端值與起始值影響。為了幫助國人對台灣茶葉與國民健康體適能的認識，我們將此演算法應用於台灣茶葉與國民健康體適能的評估系統上。根據分析的結果，本文所提之演算法確實能幫助國人對台灣茶葉與國民健康體適能進一步的認識。

In this study we propose a robust clustering method for handling LR-type fuzzy numbers. The proposed method based on the mean shift-based is not necessary to specify the cluster number and initials. Several numerical examples demonstrate the effectiveness of the proposed clustering method, especially robust to noise, outliers and initial guess. Finally, we apply this algorithm to two real data sets. One is Taiwanese tea, the other is Taiwanese residents' health-related fitness. The proposed method successfully establishes the performance evaluation system to help us better understand Taiwanese tea and residents' health-related fitness.

[1] R.O Duda, P.E. Hart, Pattern Classification and Scene Analysis, Wiley, New York, 1973.
[2] K.Fukunaga, L.D.Hostetler, The estimation of gradient of a density function with applications in pattern recognition, IEEE Trans. Inf. Theory 21 (1975) 32-40.
[3] Wen-Liang Hung, Miin-Shen Yang, Fuzzy clustering on LR-type fuzzy numbers with an application in Taiwanese tea evaluation, Fuzzy sets and systems 150(2005) 561-577.
[4] Kuo-Lung Wu, Miin-Shen Yang, Mean shift-based clustering, Pattern Recognition 40(2007) 3035-3052