TGSOM:一種用于數(shù)據(jù)聚類的動(dòng)態(tài)自組織映射神經(jīng)網(wǎng)絡(luò)
Tgsom: a new dynamic self-organizing maps for data clustering
-
摘要: 針對(duì)傳統(tǒng)Kohonen自組織特征映射(SOFM)神經(jīng)網(wǎng)絡(luò)模型結(jié)構(gòu)需預(yù)先指定的限制,提出一種新的樹(shù)形動(dòng)態(tài)自組織映射(TGSOM)神經(jīng)網(wǎng)絡(luò),當(dāng)用于數(shù)據(jù)挖掘時(shí)該網(wǎng)絡(luò)以其生成速度快可視性好具有顯著優(yōu)越性。該文詳盡描述了該網(wǎng)絡(luò)模型的生成算法,研究了算法中擴(kuò)展因子的作用。擴(kuò)展因子與訓(xùn)練樣本數(shù)據(jù)的維數(shù)無(wú)關(guān),其作用是控制網(wǎng)絡(luò)的生長(zhǎng),擴(kuò)展因子可以反映數(shù)據(jù)聚類的精度,即擴(kuò)展因子值的大小與聚類精度的高低成正比。在聚類的不同階段使用大小不等的擴(kuò)展因子還可以實(shí)現(xiàn)層次聚類。Abstract: A Tree-structured Growing Self-Organizing Maps (TGSOM) is presented as an extended version of the Self-Organizing Feature Maps (SOFM), which has significant advantages for data mining applications. The TGSOM algorithm is presented in detail and the effect of a spread factor, which can be used to measure and control the spread of the TGSOM, is investigated. The spread factor is independent of the dimensionality of the data and as such can be used as a controlling measure for generating maps with different dimensionality, which can then be compared and analyzed with better accuracy. The spread factor is also presented as a method of achieving hierarchical clustering of a data set with the TGSOM. Such hierarchical clustering allows the data analyst to identify significant and interesting clusters at a higher level of the hierarchy, and as such continue with finer clustering of only the interesting clusters.
-
M.S. Chen, J. Han, P. S. Yu, Data niining, An overview fiom a database perspective, IEEE Trans.[J]. on Knowledge Data Engineering.1996,8(6):866-[2]T. Kohonen, Self-Organization and Associate Memory, Berlin, Springer-Verlag, 1984, Chapter 5.[3]D. Alahakoon, S. K. Halgamuge, Dynamic self-organizing maps with controlled growth for knowledge discovery, IEEE Trans. on Neural Networks, 2000, NN-11(3), 601-614.[4]D. Choi, S. Park, Self-creating and organizing neural networks, IEEE Trans. on Neural Networks,1994, NN-5(4), 561-575. -
計(jì)量
- 文章訪問(wèn)數(shù): 2834
- HTML全文瀏覽量: 163
- PDF下載量: 696
- 被引次數(shù): 0
下載:
