Title :
Approaches to cluster validity index via mahalanobis metric
Author :
Yan Ren ; Lidong Wang ; Wei Guan
Author_Institution :
Coll. of Autom., Shenyang Aerosp. Univ., Shenyang, China
Abstract :
DBCAMM is a novel density based clustering algorithm via using the Mahalanobis metric, which can extract the traditional clustering information and the intrinsic clustering structure. However, one of the most significant further work of DBCAMM is to develop a cluster validity index which can indicate how to select the parameters in the algorithm. Thus, new cluster validity index IRY is proposed via using the Mahalanobis metric for the validation of partitions of object data produced by DBCAMM algorithm. The proposed index is tested and validated using several synthetic datasets with arbitrary shape. The results of the comparisons show the superior effectiveness and reliability of the proposed index in comparison to the results of other cluster validity index for FCM clustering algorithm.
Keywords :
data handling; pattern clustering; DBCAMM; Mahalanobis metric; cluster validity index; clustering information; density based clustering algorithm; intrinsic clustering structure; object data; Clustering algorithms; Data mining; Electronic mail; Indexes; Measurement; Partitioning algorithms; Shape; Arbitrary Shape Dataset; Cluster Validity Index; DBCAMM; Mahalanobis Distance;
Conference_Titel :
Control and Decision Conference (CCDC), 2015 27th Chinese
Conference_Location :
Qingdao
Print_ISBN :
978-1-4799-7016-2
DOI :
10.1109/CCDC.2015.7161986