Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Adaptive FIR Neural Model for Centroid Learning in Self-Organizing.

Slides:



Advertisements
Similar presentations
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 A 24-h forecast of solar irradiance using artificial neural.
Advertisements

Intelligent Database Systems Lab Advisor : Dr.Hsu Graduate : Keng-Wei Chang Author : Gianfranco Chicco, Roberto Napoli Federico Piglione, Petru Postolache.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Clustering data in an uncertain environment using an artificial.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A novel document similarity measure based on earth mover’s.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 SCAN: A Structural Clustering Algorithm for Networks Xiaowei.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Discovering Leaders from Community Actions Presenter : Wu, Jia-Hao Authors : Amit Goyal, Francesco Bonchi,
1 Abstract This paper presents a novel modification to the classical Competitive Learning (CL) by adding a dynamic branching mechanism to neural networks.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Fast exact k nearest neighbors search using an orthogonal search tree Presenter : Chun-Ping Wu Authors.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Unsupervised pattern recognition models for mixed feature-type.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Advisor : Dr. Hsu Student : Sheng-Hsuan Wang Department.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology U*F clustering : a new performant “ clustering-mining ”
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Human eye sclera detection and tracking using a modified.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology On multidimensional scaling and the embedding of self-organizing.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 On-line Learning of Sequence Data Based on Self-Organizing.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Graph self-organizing maps for cyclic and unbounded graphs.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Adaptive nonlinear manifolds and their applications to pattern.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 A Comparison of SOM Based Document Categorization Systems.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Exploiting Data Topology in Visualization and Clustering.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Looking inside self-organizing map ensembles with resampling.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Extracting meaningful labels for WEBSOM text archives Advisor.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Topology Preservation in Self-Organizing Feature Maps: Exact.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A self-organizing neural network using ideas from the immune.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Virus Pattern Recognition Using Self-Organization Map.
Intelligent Database Systems Lab Advisor : Dr. Hsu Graduate : Chien-Ming Hsiao Author : Bing Liu Yiyuan Xia Philp S. Yu 國立雲林科技大學 National Yunlin University.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 New Unsupervised Clustering Algorithm for Large Datasets.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Automatic Recommendations for E-Learning Personalization.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. An IPC-based vector space model for patent retrieval Presenter: Jun-Yi Wu Authors: Yen-Liang Chen, Yu-Ting.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 A Hybrid Supervised ANN for Classification and Data Visualization.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 AC-ViSOM: Hybridising the Modified Adaptive Coordinate.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Exploiting Data Topology in Visualization and Clustering.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A k-mean clustering algorithm for mixed numeric and categorical.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Extensions of vector quantization for incremental clustering.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. TurSOM: A Turing Inspired Self-organizing Map Presenter: Tsai Tzung Ruei Authors: Derek Beaton, Iren.
國立雲林科技大學 National Yunlin University of Science and Technology Self-organizing map learning nonlinearly embedded manifoldsmanifolds Author :Timo Simila.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 The Evolving Tree — Analysis and Applications Advisor.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Utilizing Marginal Net Utility for Recommendation in E-commerce.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Efficient Optimal Linear Boosting of a Pair of Classifiers.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Extensions of vector quantization for incremental clustering.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A modified version of the K-means algorithm with a distance.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. The application of SOM as a decision support tool to identify AACSB peer schools Presenter : Chun-Ping.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Advisor : Dr. Hsu Graduate : Sheng-Hsuan Wang Authors :
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Model-based evaluation of clustering validation measures.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Rival-Model Penalized Self-Organizing Map Yiu-ming Cheung.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Extending the Growing Hierarchal SOM for Clustering Documents.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Unsupervised word sense disambiguation for Korean through the acyclic weighted digraph using corpus and.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Regularization in Matrix Relevance Learning Petra Schneider,
Intelligent Database Systems Lab N.Y.U.S.T. I. M. 1 Visualization of multi-algorithm clustering for better economic decisions - The case of car pricing.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Information Loss of the Mahalanobis Distance in High Dimensions-
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Mining massive document collections by the WEBSOM method Presenter : Yu-hui Huang Authors :Krista Lagus,
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Multiclass boosting with repartitioning Graduate : Chen,
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 An initialization method to simultaneously find initial.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology O( ㏒ 2 M) Self-Organizing Map Algorithm Without Learning.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Enhanced neural gas network for prototype-based clustering.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Validity index for clusters of different sizes and densities Presenter: Jun-Yi Wu Authors: Krista Rizman.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 A self-organizing map for adaptive processing of structured.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A new data clustering approach- Generalized cellular automata.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Direct mining of discriminative patterns for classifying.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Growing Mechanisms and Cluster Identification with TurSOM.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Self Organizing Maps and Bit Signature: a study applied.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Predicting corporate bankruptcy using a self-organizing map: An empirical study to improve the forecasting.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Hierarchical model-based clustering of large datasets.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Growing Hierarchical Tree SOM: An unsupervised neural.
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Community self-Organizing Map and its Application to Data Extraction Presenter: Chun-Ping Wu Authors:
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Dual clustering : integrating data clustering over optimization.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Text Classification, Business Intelligence, and Interactivity:
Intelligent Database Systems Lab N.Y.U.S.T. I. M. Visualizing social network concepts Presenter : Chun-Ping Wu Authors :Bin Zhu, Stephanie Watts, Hsinchun.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Advisor : Dr. Hsu Graduate : Yu Cheng Chen Author: Lynette.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A Nonlinear Mapping for Data Structure Analysis John W.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 A New Cluster Validity Index for Data with Merged Clusters.
Presentation transcript:

Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Adaptive FIR Neural Model for Centroid Learning in Self-Organizing Maps Mauro Tucci and Marco Raugi TNN, 2010 Presented by Wen-Chung Liao 2010/07/28

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 2 Outlines  Motivation  Objectives  Methodology  Model Analysis  The Σ-matrix Visualization Tool  Conclusions  Comments

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 3 Motivation  There are two different SOM algorithms: the sequential algorithm, and the batch algorithm.  The batch algorithm performs one order of magnitude faster  however it is not well suited to variants and generalizations as the sequential one

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 4 Objectives  A model of the SOM processing unit for the learning of static distributions is presented.  each neuron ─ a general filter model ─ a finite impulse response (FIR) system ─ optimize the parameters of the filter in order to minimize a cost function during the training. model vector FIR coefficients trace matrix

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 5 Objectives  The FIR process of each neuron tends to become a moving average filter. ─ gives an insight of the update rule of the classic SOM algorithm. ─ improves the convergence properties with respect to the classic SOM. ─ uses a neighborhood function with a simplified design, where the annealing scheme for the learning rate is not needed. ─ used to visualize a set of properties of the input data set

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 6 Methodology N: the order of the FIR filters

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 7 Methodology D m : a distortion measure of the SOM LMS algorithm LME algorithm Step size: α

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 8 FIR-SOM Complete Learning Algorithm 1) Given the vector input data set to analyze Δ R n, create the output grid array of a finite number of cells i=1, …, D. 2) Design a decreasing function for the neighborhood width σ(t), and choose the step size α, of the filter estimator.

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 9 FIR-SOM Complete Learning Algorithm 3) For each cell, initialize the trace matrix by using random or linear initialization. 4) Choose the order of the FIR filters and initialize the coefficients to zero 5) At each time step t=0, 1, 2, …, compute the model vector of each cell i=1, …, D as

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 10 FIR-SOM Complete Learning Algorithm 6) Pick at random one sample from the input data set Δ and find the BMU as where 7) Compute the filter coefficients, for i=1, …, D, with 8) Update the trace matrix for each cell i=1, …, D, with where 9) Increase the time step and return to 5), or stop if the maximum number of iterations has been reached. (LMS algorithm)

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 11 Model Analysis A.FIR Coefficients Initialized by Zero Converge to 1/N Δ 1 : distributed in gray square 2-D 30x30 map N=10 T=40000

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 12 B. The Moving Average SOM MA-SOM C. Quality Indexes The topographic error TE(Δ) D. Convergence Properties of the MA-SOM

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 13 MA-SOM shows better quantization errors of the basic SOM for the same training duration Δ 2 : a Gaussian distribution N=10 the centroid neural network (CNN) A practical algorithm for the computation of an optimal set of unordered centroids in multivariate data. provides a guarantee of convergence to a local minimum of the quantization error. CNN required higher computational times than MA-SOM

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 14 THE Σ-MATRIX VISUALIZATION TOOL  Σ-matrix ─ ε i assumes higher values in correspondence of high- density zones  U-matrix ─ visualizing in each cell the mean of the distances between the model vector of the cell and those of the adjacent units,  P-matrix ─ based on Pareto density estimation (PDE) method  RD-matrix LME algorithm

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 15 U-matrixRD-matrix P-matrix Σ-matrix Δ 3 R 6 Two clusters, Gaussian distributions 2-D 30x30 map N=10 LME algorithm with the FIR coefficients initialized to 1/N T=40000

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 16 Δ 4 R 30 Two clusters, one Gaussian distribution, one uniform distribution 2-D 30x30 map LME algorithm N=10 P-matrix Σ-matrix

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 17 U-matrix MA-SOM,Σ-matrix WINE 178 labeled instances 13 attributes 3 types of wines 2-D 30x30 map N=10 LME algorithm Σ -matrix P-matrix

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 18 Conclusions  a good alternative to the classic SOM algorithm.  a reduced number of input presentations to reach a final state with improved map quality measures with respect to the classic SOM.  requires an added amount of basic operations and memory, but a shorter time duration of the training with respect to the classic SOM  the proposed neuron model is based on an adaptive structure, while in the classic SOM and other SOM variants, the neuron model is defined a priori.  the optimal FIR filters are moving average filters.  a proposed visualization technique, called Σ-matrix, which is based on the optimized FIR parameters.

Intelligent Database Systems Lab N.Y.U.S.T. I. M. 19 Comments  Advantage ─ Good mapping quality ─ Good visualization tool  Shortage ─ The definition of model vector ’ s adaptation is a little ambiguous.  Applications ─ Clustering ─ Classification