2004, Volume 6, Issue 9
Strategic Study of CAE >> 2004, Volume 6, Issue 9
Robust Maximum Entropy Clustering Algorithm RMEC and Its Outlier Labeling
1. School of Information Engineering , Southern Yangtze University , Wuxi , Jiangsu 214036 , China
2. National Key Lab . Of Novel Software Technologies at Nanjing University , Nanjing 210016 , China
3. School of Automation , National Defense University of Science and Technology , Changsha 410073 , China
4. Dept. Computer , Nanjing University of Science and Technology , Nanjing 212000 , China
Next Previous
Abstract
In this paper, the novel robust maximum entropy clustering algorithm RMEC, as the improved version of the maximum entropy algorithm MEC, is presented to overcome its drawbacks: very sensitive to outliers and uneasy to label them. With the introduction of Vapnik's ε-insensitive loss function and the new weight factors, the new objective function is re-constructed, and consequently, its new update rules are derived according to the Lagrangian optimization theory. Compared with algorithm MEC, the main contributions of algorithm RMEC exist in its much better robustness for outliers and the fact that it can effectively label outliers in the dataset using the obtained weight factors. The experimental results demonstrate its superior performance in enhancing the robustness and labeling outliers in the dataset.
Keywords
entropy ; clustering ; robustness ; outliers ; ε-insensitive loss function ; weight factors
Figures
图1
图2
图3
图4
图5
References
[ 1 ] RoseK , GurewtizE , FoxG Adeterministicannealingapproachtoclustering[J ].PatternRecognitionLetters, 1990, 11:589～594
[ 2 ] 邓赵红, 陆介平, 王士同.改进的MinMax模糊神经网络与函数建模[J].江南大学学报, 2003, 2 (3) :234～239 link1
[ 3 ] KellerA .Fuzzyclusteringwithoutliers[A ].NAFIPS00[M].2000
[ 4 ] KarayiannisNB .MECA :maximumentropyclusteringalgorithm[A].ProconIEEEIntConfonFuzzySyst[C].Orlando, FL , 1994.630～635
[ 5 ] LiRP , MukaidonoM .Amaximumentropyapproachtofuzzyclustering[A].ProconIEEEIntConfFuzzySyst[C].Yokohama, Japan, 1995.2227～2232
[ 6 ] 张志华, 郑南宁, 史 罡.极大熵聚类算法及其全 局收敛性分析[J].中国科学, E辑, 2001, 31 (1) :59～70 link1
[ 7 ] LasM , KandelA .Automated perceptionsindatamining[A].ProceedingsoftheEighthInternationalConferenceonFuzzySystem[C ].Seoul, Korea, 1999.190～197
[ 8 ] MendenhallW , ReinmuthJE , BeaverRJ , Statisticsformanagementandeconomics[M].Belmont, CA :DuxburyPress, 1993
[ 9 ] HuberPJ, Robuststatistics[M].NewYork:Wiley, 1981
[10] GillPE , MurrayW .WrightMH , PracticalOptimization[M].NewYork:AcademicPress, 1981
[11] 王士同.神经模糊系统及其应用[M ].北京:北京航空航天大学出版社, 1998 link1
[12] SteveRG , Supportvectormachinesclassificationandregression[R].UniversityofSouthampton, 1998
[13] VapnikV , Statisticallearningtheory[M ].NewYork:Wiley, 1998
[14] LeskiJ, Towardsarobustfuzzyclustering[J].FuzzySetsandSystems, 2003, (2) , 215～233
京公网安备 11010502051620号
