Clustering Irregular Shapes Using High-Order Neurons

This article introduces a method for clustering irregularly shaped data arrangements using high-order neurons. Complex analytical shapes are modeled by replacing the classic synaptic weight of the neuron by high-order tensors in homogeneous coordinates. In the first- and second-order cases, this neuron corresponds to a classic neuron and to an ellipsoidal-metric neuron. We show how high-order shapes can be formulated to follow the maximum-correlation activation principle and permit simple local Hebbian learning. We also demonstrate decomposition of spatial arrangements of data clusters, including very close and partially overlapping clusters, which are difficult to distinguish using classic neurons. Superior results are obtained for the Iris data.

[1]  Shigeo Abe,et al.  A fuzzy classifier with ellipsoidal regions , 1997, IEEE Trans. Fuzzy Syst..

[2]  D. Ledbetter,et al.  Multicolor Spectral Karyotyping of Human Chromosomes , 1996, Science.

[3]  Aldo Balestrino,et al.  A new adaptive polynomial neural network , 1994 .

[4]  Richard O. Duda,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[5]  G. McLachlan,et al.  The EM algorithm and extensions , 1996 .

[6]  Hichem Frigui,et al.  Fuzzy and possibilistic shell clustering algorithms and their application to boundary detection and surface approximation. II , 1995, IEEE Trans. Fuzzy Syst..

[7]  Ramanathan Gnanadesikan,et al.  Methods for statistical data analysis of multivariate observations , 1977, A Wiley publication in applied statistics.

[8]  Jeng-Shyang Pan,et al.  Fast clustering algorithms for vector quantization , 1996, Pattern Recognit..

[9]  William B. Levy,et al.  Are higher order statistics better for maximum entropy prediction in neural networks? , 1991, IJCNN-91-Seattle International Joint Conference on Neural Networks.

[10]  James C. Bezdek,et al.  Generalized clustering networks and Kohonen's self-organizing scheme , 1993, IEEE Trans. Neural Networks.

[11]  John F. Kolen,et al.  Backpropagation is Sensitive to Initial Conditions , 1990, Complex Syst..

[12]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[13]  Peter E. Hart,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[14]  Rajesh N. Dave,et al.  Use Of The Adaptive Fuzzy Clustering Algorithm To Detect Lines In Digital Images , 1990, Other Conferences.

[15]  S. Hyakin,et al.  Neural Networks: A Comprehensive Foundation , 1994 .

[16]  Anil K. Jain,et al.  A self-organizing network for hyperellipsoidal clustering (HEC) , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[17]  C. Lee Giles,et al.  Encoding Geometric Invariances in Higher-Order Neural Networks , 1987, NIPS.

[18]  Hichem Frigui,et al.  A comparison of fuzzy shell-clustering methods for the detection of ellipses , 1996, IEEE Trans. Fuzzy Syst..

[19]  S. N. Kavuri,et al.  Using fuzzy clustering with ellipsoidal units in neural networks for robust fault classification , 1993 .

[20]  I. Faux,et al.  Computational Geometry for Design and Manufacture , 1979 .

[21]  Srimat T. Chakradhar,et al.  First-order versus second-order single-layer recurrent neural networks , 1994, IEEE Trans. Neural Networks.

[22]  Alexander Graham,et al.  Kronecker Products and Matrix Calculus: With Applications , 1981 .

[23]  Donald Gustafson,et al.  Fuzzy clustering with a fuzzy covariance matrix , 1978, 1978 IEEE Conference on Decision and Control including the 17th Symposium on Adaptive Processes.

[24]  Blatt,et al.  Superparamagnetic clustering of data. , 1998, Physical review letters.