• ETRI Journal
• /
• 제32권1호
• /
• pp.160-162
• /
• 2010

For the acoustic models of embedded speech recognition systems, hidden Markov models (HMMs) are usually quantized and the original full space distributions are represented by combinations of a few quantized distribution prototypes. We propose a maximum likelihood objective function to train the quantized distribution prototypes. The experimental results show that the new training algorithm and the link structure adaptation scheme for the quantized HMMs reduce the word recognition error rate by 20.0%.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제50권7호
• /
• pp.339-349
• /
• 2001

This paper suggests an optimal identification method for complex and nonlinear system modeling that is based on Fuzzy-Neural Networks(FNN). The proposed Hybrid Identification Algorithm is based on Yamakawa's FNN and uses the simplified inference as fuzzy inference method and Error Back Propagation Algorithm as learning rule. In this paper, the FNN modeling implements parameter identification using HCM algorithm and hybrid structure combined with two types of optimization theories for nonlinear systems. We use a HCM(Hard C-Means) clustering algorithm to find initial apexes of membership function. The parameters such as apexes of membership functions, learning rates, and momentum coefficients are adjusted using hybrid algorithm. The proposed hybrid identification algorithm is carried out using both a genetic algorithm and the improved complex method. Also, an aggregated objective function(performance index) with weighting factor is introduced to achieve a sound balance between approximation and generalization abilities of the model. According to the selection and adjustment of a weighting factor of an aggregate objective function which depends on the number of data and a certain degree of nonlinearity(distribution of I/O data), we show that it is available and effective to design an optimal FNN model structure with mutual balance and dependency between approximation and generalization abilities. To evaluate the performance of the proposed model, we use the time series data for gas furnace, the data of sewage treatment process and traffic route choice process.

• Geomechanics and Engineering
• /
• 제11권4호
• /
• pp.493-513
• /
• 2016

A modified grey clustering method is presented to systematically evaluate the risk of water inrush in karst tunnels. Based on the center triangle whitenization weight function and upper and lower limit measure whitenization weight function, the modified grey evaluation model doesn't have the crossing properties of grey cluster and meets the standard well. By adsorbing and integrating the previous research results, seven influence factors are selected as evaluation indexes. A couple of evaluation indexes are modified and quantitatively graded according to four risk grades through expert evaluation method. The weights of evaluation indexes are rationally distributed by the comprehensive assignment method. It is integrated by the subjective factors and the objective factors. Subjective weight is given based on analytical hierarchy process, and objective weight obtained from simple dependent function. The modified grey evaluation model is validated by Jigongling Tunnel. Finally, the water inrush risk of Shangjiawan Tunnel is evaluated by using the established model, and the evaluation result obtained from the proposed method is agrees well with practical situation. This risk assessment methodology provides a powerful tool with which planners and engineers can systematically assess the risk of water inrush in karst tunnels.

• Journal of Electrical Engineering and Technology
• /
• 제6권6호
• /
• pp.853-866
• /
• 2011

We propose a multi-objective space search algorithm (MSSA) and introduce the identification of fuzzy inference systems based on the MSSA and information granulation (IG). The MSSA is a multi-objective optimization algorithm whose search method is associated with the analysis of the solution space. The multi-objective mechanism of MSSA is realized using a non-dominated sorting-based multi-objective strategy. In the identification of the fuzzy inference system, the MSSA is exploited to carry out parametric optimization of the fuzzy model and to achieve its structural optimization. The granulation of information is attained using the C-Means clustering algorithm. The overall optimization of fuzzy inference systems comes in the form of two identification mechanisms: structure identification (such as the number of input variables to be used, a specific subset of input variables, the number of membership functions, and the polynomial type) and parameter identification (viz. the apexes of membership function). The structure identification is developed by the MSSA and C-Means, whereas the parameter identification is realized via the MSSA and least squares method. The evaluation of the performance of the proposed model was conducted using three representative numerical examples such as gas furnace, NOx emission process data, and Mackey-Glass time series. The proposed model was also compared with the quality of some "conventional" fuzzy models encountered in the literature.

• ETRI Journal
• /
• 제23권4호
• /
• pp.168-176
• /
• 2001

This paper discusses two important issues of corpus-based synthesis: synthesis unit generation based on phrase break strength information and pruning redundant synthesis unit instances. First, the new sentence set for recording was designed to make an efficient synthesis database, reflecting the characteristics of the Korean language. To obtain prosodic context sensitive units, we graded major prosodic phrases into 5 distinctive levels according to pause length and then discriminated intra-word triphones using the levels. Using the synthesis unit with phrase break strength information, synthetic speech was generated and evaluated subjectively. Second, a new pruning method based on weighted vector quantization (WVQ) was proposed to eliminate redundant synthesis unit instances from the synthesis database. WVQ takes the relative importance of each instance into account when clustering similar instances using vector quantization (VQ) technique. The proposed method was compared with two conventional pruning methods through objective and subjective evaluations of synthetic speech quality: one to simply limit the maximum number of instances, and the other based on normal VQ-based clustering. For the same reduction rate of instance number, the proposed method showed the best performance. The synthetic speech with reduction rate 45% had almost no perceptible degradation as compared to the synthetic speech without instance reduction.

• 산업공학
• /
• 제15권2호
• /
• pp.114-125
• /
• 2002

The objective of this paper is to describe a systematic accident scenario analysis method(SASA) adept at creating accident scenarios for the design of safer products. This approach was inspired by the Quality Function Deployment(QFD) method, which is conventionally used in quality management. In this study, the QFD provides a formal and systematic scheme to devise accident scenarios while maintaining objectivity. SASA consists of three key stages to be broken down into a series of consecutive steps:(1) developing an accident analysis tableau,(2) devising the accident scenarios using the accident analysis tableau,(3) performing a feasibility test, a clustering process and a patterning process, and finally(4) performing quantitative evaluation of each accident scenario. The SASA was applied to a case study of child safety seats. The accident analysis tableau devised 2828(maximum) accident scenarios from all possible relationships between the hazard factors and situation characteristics. Among them, 270 scenarios were devised through the feasibility test and the clustering process. The patterning process reduced them to 29 patterns representative of all accident scenarios. Based on an intensive analysis of the accident patterns, design guidelines for a safer child safety seat were recommended. The implications of the study on the child safety seat case were then discussed.

• 한국공작기계학회논문집
• /
• 제15권5호
• /
• pp.104-113
• /
• 2006

The main objective of this study is to investigate the effect of compounded welding by using acoustic emission (AE) signals and doing a source location for weld heat affected zone (HAZ) through tensile testing. This study was carried out an SWS 490A high strength steel for electric shield metal arc welding, SMAW; $CO_2$ gas metal arc welding, GMAW($CO_2$); and gas tungsten arc welding, GTAW/TIG. Data displays are based on the measured parameters of the AE signals, along with environmental variables such as time and load. For instance, Gutenberg-Richter magnitude-frequency relationship (G-R MFR) offers useful b-value in data analysis. Namely event identification, source location gives the X- and Y-coordinates of the AE source. And K-means clustering analysis by Euclidean distance confirmed that was powerful to source location. Generally, strength of welded metal zone was stronger than strength of base metal. As the result, confirmed certainly that fracture is produced in HAZ instead of welded metal zone from source location.

• 경영과학
• /
• 제16권1호
• /
• pp.157-171
• /
• 1999

In a increasingly competitive marketplace, the manufacturing companies have no choice but looking for ways to improve productivity to sustain their competitiveness and survive in the industry. Recently cellular manufacturing has been under discussion as an option to be easily implemented without burdensome capital investment. The objective of cellular manufacturing is to realize many aspects of efficiencies associated with mass production in the less repetitive job-shop production systems. The very first step for cellular manufacturing is to group the sets of parts having similar processing requirements into part families, and the equipment needed to process a particular part family into machine cells. The underlying problem to determine the part and machine assignments to each manufacturing cell is called the cell formation. The purpose of this study is to develop a clustering algorithm based on the neural network approach which overcomes the drawbacks of ART1 algorithm for cell formation problems. In this paper, a generalized learning vector quantization(GLVQ) algorithm was devised in order to transform a 0/1 part-machine assignment matrix into the matrix with diagonal blocks in such a way to increase clustering performance. Furthermore, an assignment problem model and a rearrangement procedure has been embedded to increase efficiency. The performance of the proposed algorithm has been evaluated using data sets adopted by prior studies on cell formation. The proposed algorithm dominates almost all the cell formation reported so far, based on the grouping index($\alpha$ = 0.2). Among 27 cell formation problems investigated, the result by the proposed algorithm was superior in 11, equal 15, and inferior only in 1.

• 대한기계학회논문집A
• /
• 제25권11호
• /
• pp.1796-1801
• /
• 2001

For micro-machines, very few design methodologies based on optimization hale been developed so far. To overcome the difficulties of design optimization of micro-machines, the search method for multi-dimensional design window (DW)s is proposed. The proposed method is defined as areas of satisfactory design solutions in a design parameter space, using both continuous evolutionary algorithms (CEA) and the modified K-means clustering algorithm . To demonstrate practical performance of the proposed method, it was applied to an optimal shape design of micro electrostatic actuator of optical memory. The shape design problem has 5 design parameters and 5 objective functions, and finally shows 4 specific design shapes and design characters based on the proposed DWs.

• 제어로봇시스템학회논문지
• /
• 제6권3호
• /
• pp.273-283
• /
• 2000

In this paper we suggest an optimal design method of Fuzzy-Neural Networks(FNN) model for complex and nonlinear systems. The FNNs use the simplified inference as fuzzy inference method and Error Back Propagation Algorithm as learning rule. And we use a HCM(Hard C-Means) Clustering Algorithm to find initial parameters of the membership function. The parameters such as parameters of membership functions learning rates and momentum weighted value is proposed to achieve a sound balance between approximation and generalization abilities of the model. According to selection and adjustment of a weighting factor of an aggregate objective function which depends on the number of data and a certain degree of nonlinearity (distribution of I/O data we show that it is available and effective to design and optimal FNN model structure with a mutual balance and dependency between approximation and generalization abilities. This methodology sheds light on the role and impact of different parameters of the model on its performance (especially the mapping and predicting capabilities of the rule based computing). To evaluate the performance of the proposed model we use the time series data for gas furnace the data of sewage treatment process and traffic route choice process.