How to efficiently specify the “correct” number of clusters from a given multidimensional data se... more How to efficiently specify the “correct” number of clusters from a given multidimensional data set is one of the most fundamental and unsolved problems in cluster analysis. In this paper, we propose a method for automatically discovering the number of clusters and estimating the locations of the centroids of the resulting clusters. This method is based on the interpretation of
A new approach to data clustering which is capable of detecting clusters of different shapes is p... more A new approach to data clustering which is capable of detecting clusters of different shapes is proposed. In classical clustering approaches, it is a great challenge to separate clusters if the cluster prototypes are difficult to represent by a mathematical formula. In this paper, we propose an improved learning vector quantization (LVQ) algorithm using the concept of symmetry. Through several computer simulations, the results show that the proposed method with random initialization is effective in detecting linear, spherical and ellipsoidal clusters. Besides, this method can also solve the crossed question
For reducing requirement of large memory and minimizing computation complexity in large-vocabular... more For reducing requirement of large memory and minimizing computation complexity in large-vocabulary continuous speech recognition system, speech segmentation plays an important role in speech recognition systems. In this paper, we formulate the speech segmentation as a twephase problem. Phase 1 (frame labelling) involves labeling frames of speech data. Frames are classified into three types : (1) silence; (2) consonants and (3) vowels according to two segmentation features. In phase 2 (syllabic unit segmentation) we apply the concept of transition states to segment continuous speech data into syllabic units based on the labeled frames. The novel class of hyperrectangular composite neural networks (HRCNN's) is used to cluster frames. The HRCNN's integrate the rule-based approach and neural network paradigms, therefore, this special hybrid system may neutralize the disadvantages of each alternative. The parameters in the trained HRCNN's are utilized to extract both crispy and fuzzy classification rules. Four speaker's continuous reading-rate Mandarin speech are given to illustrate the proposed two-phase speech segmentation model. In our experiments, the performance of the HRCNN's is better than the "Distributed Fuzzy Rule" approach based on the comparisons of the number of rules and the correct recognition rate.
This paper presents a neuro-fuzzy system by using the Kohonen 's self-organizing feature map... more This paper presents a neuro-fuzzy system by using the Kohonen 's self-organizing feature map algorithm, not only for its vector yuantitization feature, but also for its topological property. This property prevents the proposed neuro-fuzzy system from suffering from a drawback like ...
This paper presents a hierarchical neuro-fuzzy network system for segmenting continuous speech in... more This paper presents a hierarchical neuro-fuzzy network system for segmenting continuous speech into syllables. The system formulates speech segmentation as a two-phase procedure. In the first phase, the Hybrid Neuro-Fuzzy network (HNFN) is utilized to classify the speech signal into three different types. The hybrid model (HNFN) composed of a distributed representation of a fuzzy system (DRF) and a hyperrectangular composite neural network (HRCNN) is proposed and used to cluster frames. This special hybrid system may neutralize the disadvantages of each alternative. The parameters in the trained HNFN are utilized to extract both crispy and fuzzy classification rules. In the following phase, the self-tuning back-propagation neural network (STBNN) is utilized to solve the coarticulation effects of vowel-vowel (V-V) concatenation. In our experiments, a database containing continuous reading-rate Mandarin speech recorded from newscasts was utilized to test the performance of the propose...
In this chapter we present a method for the identification of fuzzy singleton models based on a c... more In this chapter we present a method for the identification of fuzzy singleton models based on a class of Fuzzy HyperRectangular Composite Neural Networks (FHRCNNs). The prior knowledge required, besides the available input-output data, is knowledge about the input and output variables and the number of partitions dividing the ranges of the output variables.
1997 IEEE International Conference on Systems, Man, and Cybernetics. Computational Cybernetics and Simulation
How to efficiently specify the “correct” number of clusters from a given multidimensional data se... more How to efficiently specify the “correct” number of clusters from a given multidimensional data set is one of the most fundamental and unsolved problems in cluster analysis. In this paper, we propose a method for automatically discovering the number of clusters and estimating the locations of the centroids of the resulting clusters. This method is based on the interpretation of
A fuzzy ARTMAP system is a system for incremental supervised learning of recognition categories a... more A fuzzy ARTMAP system is a system for incremental supervised learning of recognition categories and multidimensional maps in response to arbitrary sequence of analog or binary input vectors. The original fuzzy ARTMAP system incorporates two fuzzy ART modules and an inter-ART module. Many different approaches have been proposed to modify fuzzy ARTMAP systems. In this paper, we proposed a new approach to modifying a fuzzy ARTMAP system. We referred to the new system as the Modified and Simplified Fuzzy ARTMAP (MSFAM) system. Two data sets were used for demonstrating the performance of the proposed MSFAM systems. Keyword: neural networks, learning algorithm, ART
Recently, feature maps have been applied to various problem domains. The success of some of these... more Recently, feature maps have been applied to various problem domains. The success of some of these applications critically depends on whether feature maps are topologically ordered. In this paper, we propose a novel measure for quantifying the neighborhood preserving property of feature maps. Two data sets were tested to illustrate the performance of the proposed method.
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium, 2000
How to efficiently specify the “correct” number of clusters from a given multidimensional data se... more How to efficiently specify the “correct” number of clusters from a given multidimensional data set is one of the most fundamental and unsolved problems in cluster analysis. In this paper, we propose a method for automatically discovering the number of clusters and estimating the locations of the centroids of the resulting clusters. This method is based on the interpretation of
A new approach to data clustering which is capable of detecting clusters of different shapes is p... more A new approach to data clustering which is capable of detecting clusters of different shapes is proposed. In classical clustering approaches, it is a great challenge to separate clusters if the cluster prototypes are difficult to represent by a mathematical formula. In this paper, we propose an improved learning vector quantization (LVQ) algorithm using the concept of symmetry. Through several computer simulations, the results show that the proposed method with random initialization is effective in detecting linear, spherical and ellipsoidal clusters. Besides, this method can also solve the crossed question
For reducing requirement of large memory and minimizing computation complexity in large-vocabular... more For reducing requirement of large memory and minimizing computation complexity in large-vocabulary continuous speech recognition system, speech segmentation plays an important role in speech recognition systems. In this paper, we formulate the speech segmentation as a twephase problem. Phase 1 (frame labelling) involves labeling frames of speech data. Frames are classified into three types : (1) silence; (2) consonants and (3) vowels according to two segmentation features. In phase 2 (syllabic unit segmentation) we apply the concept of transition states to segment continuous speech data into syllabic units based on the labeled frames. The novel class of hyperrectangular composite neural networks (HRCNN's) is used to cluster frames. The HRCNN's integrate the rule-based approach and neural network paradigms, therefore, this special hybrid system may neutralize the disadvantages of each alternative. The parameters in the trained HRCNN's are utilized to extract both crispy and fuzzy classification rules. Four speaker's continuous reading-rate Mandarin speech are given to illustrate the proposed two-phase speech segmentation model. In our experiments, the performance of the HRCNN's is better than the "Distributed Fuzzy Rule" approach based on the comparisons of the number of rules and the correct recognition rate.
This paper presents a neuro-fuzzy system by using the Kohonen 's self-organizing feature map... more This paper presents a neuro-fuzzy system by using the Kohonen 's self-organizing feature map algorithm, not only for its vector yuantitization feature, but also for its topological property. This property prevents the proposed neuro-fuzzy system from suffering from a drawback like ...
This paper presents a hierarchical neuro-fuzzy network system for segmenting continuous speech in... more This paper presents a hierarchical neuro-fuzzy network system for segmenting continuous speech into syllables. The system formulates speech segmentation as a two-phase procedure. In the first phase, the Hybrid Neuro-Fuzzy network (HNFN) is utilized to classify the speech signal into three different types. The hybrid model (HNFN) composed of a distributed representation of a fuzzy system (DRF) and a hyperrectangular composite neural network (HRCNN) is proposed and used to cluster frames. This special hybrid system may neutralize the disadvantages of each alternative. The parameters in the trained HNFN are utilized to extract both crispy and fuzzy classification rules. In the following phase, the self-tuning back-propagation neural network (STBNN) is utilized to solve the coarticulation effects of vowel-vowel (V-V) concatenation. In our experiments, a database containing continuous reading-rate Mandarin speech recorded from newscasts was utilized to test the performance of the propose...
In this chapter we present a method for the identification of fuzzy singleton models based on a c... more In this chapter we present a method for the identification of fuzzy singleton models based on a class of Fuzzy HyperRectangular Composite Neural Networks (FHRCNNs). The prior knowledge required, besides the available input-output data, is knowledge about the input and output variables and the number of partitions dividing the ranges of the output variables.
1997 IEEE International Conference on Systems, Man, and Cybernetics. Computational Cybernetics and Simulation
How to efficiently specify the “correct” number of clusters from a given multidimensional data se... more How to efficiently specify the “correct” number of clusters from a given multidimensional data set is one of the most fundamental and unsolved problems in cluster analysis. In this paper, we propose a method for automatically discovering the number of clusters and estimating the locations of the centroids of the resulting clusters. This method is based on the interpretation of
A fuzzy ARTMAP system is a system for incremental supervised learning of recognition categories a... more A fuzzy ARTMAP system is a system for incremental supervised learning of recognition categories and multidimensional maps in response to arbitrary sequence of analog or binary input vectors. The original fuzzy ARTMAP system incorporates two fuzzy ART modules and an inter-ART module. Many different approaches have been proposed to modify fuzzy ARTMAP systems. In this paper, we proposed a new approach to modifying a fuzzy ARTMAP system. We referred to the new system as the Modified and Simplified Fuzzy ARTMAP (MSFAM) system. Two data sets were used for demonstrating the performance of the proposed MSFAM systems. Keyword: neural networks, learning algorithm, ART
Recently, feature maps have been applied to various problem domains. The success of some of these... more Recently, feature maps have been applied to various problem domains. The success of some of these applications critically depends on whether feature maps are topologically ordered. In this paper, we propose a novel measure for quantifying the neighborhood preserving property of feature maps. Two data sets were tested to illustrate the performance of the proposed method.
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium, 2000
Uploads
Papers by Mu-Chun Su