• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2002.11a
• /
• pp.382-385
• /
• 2002

We design LVQ network using SOM network for the LVQ's performance improvement. Reference vectors and the number of output neurons, they are the proposed LVQ network's initial parameters, are determined in SOM which is used for preprocessing LVQ. We simulate it to the grouping test of Fisher's Iris data. In this result, we confirm proposed LVQ network is better than existing LVQ in grouping test.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.05a
• /
• pp.245-248
• /
• 2003

In this paper, we propose a learning method of the performance improvement of the LVQ network using the radios of the hypersphere with the n-dimensional input vectors. The proposed method determines the reference vectors using the radius of the hypersphere include n+1 set of input vectors in the same class. In order to verify the effectiveness of the proposed method, we performed experiments on the Fisher's IRIS data. The experimental results showed that the proposed method improves considerably on the performance of the conventional LVQ network.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.7
• /
• pp.1291-1300
• /
• 2008

In this paper, we propose a intelligent system to improve the accuracy of pattern recognition. The proposed intelligent system consist in SOFM, LVQ and FCM algorithm. We are confirmed the effectiveness of the proposed intelligent system through the several experiments that classify Fisher's Iris data and face image data that offered by ORL of Cambridge Univ. and EMG data. As the results of experiments, the proposed intelligent system has better accuracy of pattern recognition than general LVQ.

• The Korean Journal of Applied Statistics
• /
• v.16 no.2
• /
• pp.321-333
• /
• 2003

Self-organizing map (SOM), a unsupervised learning neural network, has been developed by T. Kohonen since 1980's. Main application areas were pattern recognition and text retrieval. Because of that, it has not been spread to statisticians until late. Recently, SOM's are frequently drawn in data mining fields. Kohonen's SOM, however, needs improvements to become a statistician's standard tool. First, there should be a good guideline as for the size of map. Second, an enhanced visualization mode is wanted. In this study, principal components self-organizing map (PC-SOM), a modification of Kohonen's SOM, is proposed to meet such needs. PC-SOM performs one-dimensional SOM during the first stage to decompose input units into node weights and residuals. At the second stage, another one-dimensional SOM is applied to the residuals of the first stage. Finally, by putting together two stages, one obtains two-dimensional SOM. Such procedure can be easily expanded to construct three or more dimensional maps. The number of grid lines along the second axis is determined automatically, once that of the first axis is given by the data analyst. Furthermore, PC-SOM provides easily interpretable map axes. Such merits of PC-SOM are demonstrated with well-known Fisher's iris data and a simulated data set.

• Journal of the Korean Professional Engineers Association
• /
• v.29 no.2
• /
• pp.71-79
• /
• 1996

This paper evaluates the clustering performance of a neural network and a statistical method. Algorithms which are used in this paper are the GLVQ(Generalized Loaming vector Quantization) for a neural method and the k -means algorithm for a statistical clustering method. For comparison of two methods, we calculate the Rand's c statistics. As a result, the mean of c value obtained with the GLVQ is higher than that obtained with the k -means algorithm, while standard deviation of c value is lower. Experimental data sets were the Fisher's IRIS data and patterns extracted from handwritten numerals.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.2 s.308
• /
• pp.33-39
• /
• 2006

This paper presents the modified LVQ method for performance improvement of pattern classification. The proposed method uses the skewness of probability distribution between the input vectors and the reference vectors. During training, the reference vectors are closest to the input vectors using the probabilistic distribution of the input vectors, and they are positioned to approximate the decision surfaces of the theoretical Bayes classifier. In order to verify the effectiveness of the proposed method, we performed experiments on the Gaussian distribution data set, and the Fisher's IRIS data set. The experimental results show that the proposed method considerably improves on the performance of the LVQ1, LVQ2, and GLVQ.

• The Korean Journal of Applied Statistics
• /
• v.18 no.3
• /
• pp.661-671
• /
• 2005

The aim of this study is to propose a clustering method, called tree-structured clustering, by recursively partitioning continuous multivariate dat a based on overall $R^2$ criterion with a practical node-splitting decision rule. The clustering method produces easily interpretable clustering rules of tree types with the variable selection function. In numerical examples (Fisher's iris data and a Telecom case), we note several differences between tree-structured clustering and K-means clustering.

• Proceedings of the IEEK Conference
• /
• 2003.07d
• /
• pp.1267-1270
• /
• 2003

In this paper, we propose a optimal design method of the LVQ network. The proposed method determines the initial reference vectors and optimal network structure using the winning expectation of output neurons of SOM. In order to verify the effectiveness of the proposed method, we performed experiments on the Fisher's IRIS data. The experimental results showed that the proposed method improves considerably on the performance of the conventional LVQ networks.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.4
• /
• pp.460-465
• /
• 2007

In this paper, we propose a novel pattern classification method based on statistical properties of the given data and fuzzy-rough set to minimize the coupling problem of the rules. In the proposed method, statistical properties is used by a selection criteria for deciding a partition number of antecedent fuzzy sets, and for minimizing an coupling problem of the generated rules. Moreover, rough set is used as a tool to remove unnecessary attributes between generated rules from the numerical data. In order to verify the validity of the proposed method, we compared the classification results (i.e, classification precision) of the proposed with the conventional pattern classification methods on the Fisher's IRIS data. From experiment results, we can conclude that the proposed method shows relatively better performance than those of the classification methods based on the conventional approaches.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.6
• /
• pp.792-798
• /
• 2007

In this paper, we propose a hierarchical fuzzy-rough classification method based on statistical information for maximizing the performance of pattern classification and reducing the number of rules without learning approaches such as neural network, genetic algorithm. In the proposed method, statistical information is used for extracting the partition intervals of antecedent fuzzy sets at each layer on hierarchical fuzzy-rough classification systems and rough sets are used for minimizing the number of fuzzy if-then rules which are associated with the partition intervals extracted by statistical information. To show the effectiveness of the proposed method, we compared the classification results(e.g. the classification accuracy and the number of rules) of the proposed with those of the conventional methods on the Fisher's IRIS data. From the experimental results, we can confirm the fact that the proposed method considers only statistical information of the given data is similar to the classification performance of the conventional methods.