Improving the Fuzzy Min-Max neural network performance with an ensemble of clustering trees

Abstract In this paper, an ensemble of clustering trees (ECTs) is adopted to improve the performance of the Fuzzy Min–Max (FMM) network with individual clustering trees. The key advantage of combining FMM and ECT together is to formulate an accurate and useful learning model that is able to perform online clustering and to explain its predictions. The online clustering capability is inherited from the FMM hyperboxes, while the explanatory capability arises from the underlying decision trees of ECT. Four different mean measures, namely harmonic, geometric, arithmetic, and root mean square, are incorporated into FMM for computing its hyperbox centroids. A series of benchmark and real-world data sets are used for evaluating the FMM-ECT performance. The results are analyzed and compared with those from other models. The outcomes indicate that FMM-ECT is able to achieve comparable clustering performances, with the advantage of providing explanations of its predictions using a decision tree.

[1]  Yong Li,et al.  Integrated dynamic evaluation of depletion-drive performance in naturally fractured-vuggy carbonate reservoirs using DPSO–FCM clustering , 2016 .

[2]  C. D. Smothers,et al.  Comparison of arithmetic and geometric means as measures of a central tendency in cattle nematode populations. , 1999, Veterinary parasitology.

[3]  Bo Jiang,et al.  Bi-level weighted multi-view clustering via hybrid particle swarm optimization , 2016, Inf. Process. Manag..

[4]  Raymond J. Mooney,et al.  Combining Bias and Variance Reduction Techniques for Regression Trees , 2005, ECML.

[5]  Abdolreza Mirzaei,et al.  A hierarchical clusterer ensemble method based on boosting theory , 2013, Knowl. Based Syst..

[6]  Ekrem Duman,et al.  A profit-driven Artificial Neural Network (ANN) with applications to fraud detection and direct marketing , 2016, Neurocomputing.

[7]  Subrahmanyam Pulipaka,et al.  Power prediction of soiled PV module with neural networks using hybrid data clustering and division techniques , 2016 .

[8]  Ahmed Hamza Osman,et al.  Enhancement of spam detection mechanism based on hybrid $$\varvec{k}$$k-mean clustering and support vector machine , 2015, Soft Comput..

[9]  Friedhelm Schwenker,et al.  Selective neural network ensembles in reinforcement learning: Taking the advantage of many agents , 2015, Neurocomputing.

[10]  Abdolreza Mirzaei,et al.  A Novel Hierarchical-Clustering-Combination Scheme Based on Fuzzy-Similarity Relations , 2010, IEEE Transactions on Fuzzy Systems.

[11]  Hamit Erdal,et al.  Bagging ensemble models for bank profitability: An emprical research on Turkish development and investment banks , 2016, Appl. Soft Comput..

[12]  Yong Peng,et al.  An unsupervised discriminative extreme learning machine and its applications to data clustering , 2016, Neurocomputing.

[13]  Antanas Verikas,et al.  Mining data with random forests: A survey and results of new tests , 2011, Pattern Recognit..

[14]  Chee Peng Lim,et al.  Power Quality Analysis Using a Hybrid Model of the Fuzzy Min–Max Neural Network and Clustering Tree , 2016, IEEE Transactions on Neural Networks and Learning Systems.

[15]  Ren-long Zhang,et al.  A novel fuzzy hybrid quantum artificial immune clustering algorithm based on cloud model , 2014, Eng. Appl. Artif. Intell..

[16]  Serge Guillaume,et al.  DENDIS: A new density-based sampling for clustering algorithm , 2016, Expert Syst. Appl..

[17]  Hong Gu,et al.  A hybrid genetic algorithm–fuzzy c-means approach for incomplete data clustering based on nearest-neighbor intervals , 2013, Soft Comput..

[18]  Patrick K. Simpson,et al.  Fuzzy min-max neural networks - Part 2: Clustering , 1993, IEEE Trans. Fuzzy Syst..

[19]  Jiawei Han,et al.  Data Mining: Concepts and Techniques , 2000 .

[20]  Jian Zhou,et al.  A modified hybrid method of spatial credibilistic clustering and particle swarm optimization , 2011, Soft Comput..

[21]  F. Sheikholeslam,et al.  Fuzzy Wavelet Neural Network With an Accelerated Hybrid Learning Algorithm , 2012, IEEE Transactions on Fuzzy Systems.

[22]  Seung-Bok Choi,et al.  Hybrid clustering based fuzzy structure for vibration control - Part 1: A novel algorithm for building neuro-fuzzy system , 2015 .

[23]  Rasool Azimi,et al.  A hybrid method based on a new clustering technique and multilayer perceptron neural networks for hourly solar radiation forecasting , 2016 .

[24]  Joshua C. C. Chan,et al.  Pitfalls of Estimating the Marginal Likelihood Using the Modified Harmonic Mean , 2015 .

[25]  Sung-Kwun Oh,et al.  Design of hybrid radial basis function neural networks (HRBFNNs) realized with the aid of hybridization of fuzzy clustering method (FCM) and polynomial neural networks (PNNs) , 2014, Neural Networks.

[26]  Jun Sun,et al.  A hybrid fuzzy K-harmonic means clustering algorithm , 2015 .

[27]  Wei-Chang Yeh,et al.  A novel hybrid clustering approach based on K-harmonic means using robust design , 2016, Neurocomputing.

[28]  Yifan Zhou,et al.  Hybrid model of clustering and kernel autoassociator for reliable vehicle type classification , 2013, Machine Vision and Applications.