This paper presents an advanced fuzzy C-means (FCM) clustering algorithm to overcome the weakness
of the traditional FCM algorithm, including the instability of random selecting of initial center and the limitation
of the data separation or the size of clusters. The advanced FCM algorithm combines the distance with density
and improves the objective function so that the performance of the algorithm can be improved. The experimental
results show that the proposed FCM algorithm requires fewer iterations yet provides higher accuracy than the
traditional FCM algorithm. The advanced algorithm is applied to the influence of stars’ box-office data, and the
classification accuracy of the first class stars achieves 92.625%.
WU Shaochun (吴绍春), PANG Yijie (庞毅杰), SHAO Sen (邵森), JIANG Keyuan (江科元)
. Advanced Fuzzy C-Means Algorithm Based on Local Density and Distance[J]. Journal of Shanghai Jiaotong University(Science), 2018
, 23(5)
: 636
-642
.
DOI: 10.1007/s12204-018-1993-y
[1] JIANG L, DING H F. Improvement of the fuzzy Cmeansclustering algorithm [J]. Computer & DigitalEngineering, 2010, 38(2): 4-6 (in Chinese).
[2] WANG X,WANG C F, SHEN J Y. Semi-supervised kmeansclustering by optimizing initial cluster centers[J]. Lecture Notes in Computer Science, 2011, 6988:178-187.
[3] SUN J G, LIU J, ZHAO L Y. Clustering algorithmsresearch [J]. Journal of Software, 2008, 19(1): 48-61(in Chinese).
[4] LIU X M, LEI D. An improved fuzzy C-means clusteringalgorithm [J]. Journal of Qingdao Universityof Science and Technology (Natural Science Edition),2011, 32(2): 194-198 (in Chinese).
[5] CHEN L L, AN C L. Adaptive fuzzy C-means algorithmon density and validity [J]. Microcomputer & ItsApplications, 2012, 31(8): 194-198 (in Chinese).
[6] LI J F, ZHU F F. Improved FCM algorithm basedon the density of samples [J]. Science Technology andEngineering, 2007, 7(4): 636-638 (in Chinese).
[7] LU C H, XIAO S Q, GU X F. Improving fuzzy Cmeansclustering algorithm based on a density-induceddistance measure [J]. Journal of Engineering, 2014.https://doi.org/10.1049/joe.2014.0053 (published online).
[8] DUNN J C. A fuzzy relative of the ISODATA processand its use in detecting compact well-separated clusters[J]. Journal of Cybernetics, 1974, 3(3): 32-57.
[9] BEZDEK J C. Cluster validity with fuzzy sets [J].Journal of Cybernetics, 1974, 3(3): 58-73.
[10] LIU L F, SUN S Z, YU H Y, et al. A modified fuzzy Cmeans(FCM) clustering algorithm and its applicationon carbonate fluid identification [J]. Journal of AppliedGeophysics, 2016, 129: 28-35.
[11] YANG M S, NATALIANI Y. Robust-learning fuzzy Cmeansclustering algorithm with unknown number ofclusters [J]. Pattern Recognition, 2017, 71: 45-59.
[12] HATHAWAY R J, BEZDEK J C, HU Y K. Generalizedfuzzy C-means clustering strategies using Lp norm distances[J]. IEEE Transactions on Fuzzy Systems, 2000,8(5): 576-582.
[13] BEZDEK J C. Pattern recognition with fuzzy objectivefunction algorithms [M]. New York, USA: PlenumPress, 1981.
[14] YU F H, XU H K, WANG L M, et al. An improved automaticFCM clustering algorithm [C]//Proceedings ofInternational Workshop on Database Technology andApplications. Wuhan, China: IEEE, 2010: 1-4.
[15] QIAN Y T, ZHAO R C. Robust clustering based onglobal data distribution and local connectivity matrix[C]//IEEE International Conference on IntelligentProcessing Systems. Beijing, China: IEEE, 1997: 1629-1633.
[16] PIAO S Z, CHAOMURILIGE, YU J. Cluster validityindexes for FCM clustering algorithm [J]. PatternRecognition & Artificial Intelligence, 2015, 28(5): 452-461 (in Chinese).
