• Phonetics and Speech Sciences
• /
• v.7 no.3
• /
• pp.131-138
• /
• 2015

In this paper, we propose L1-norm regularization for state vector adaptation of subspace Gaussian mixture model (SGMM). When you design a speaker adaptation system with GMM-HMM acoustic model, MAP is the most typical technique to be considered. However, in MAP adaptation procedure, large number of parameters should be updated simultaneously. We can adopt sparse adaptation such as L1-norm regularization or sparse MAP to cope with that, but the performance of sparse adaptation is not good as MAP adaptation. However, SGMM does not suffer a lot from sparse adaptation as GMM-HMM because each Gaussian mean vector in SGMM is defined as a weighted sum of basis vectors, which is much robust to the fluctuation of parameters. Since there are only a few adaptation techniques appropriate for SGMM, our proposed method could be powerful especially when the number of adaptation data is limited. Experimental results show that error reduction rate of the proposed method is better than the result of MAP adaptation of SGMM, even with small adaptation data.

• Journal of Korean Academy of Nursing
• /
• v.53 no.4
• /
• pp.426-438
• /
• 2023

Purpose: Walking practice is a representative indicator of the level of physical activity of local residents. Although the world health organization addressed reduction in prevalence of insufficient physical activity as a global target, the rate of walking practice in Korea has not improved and there are large regional disparities. Therefore, this study aimed to explore the spatial variations of walking practice and its associated factors in Korea. Methods: A secondary analysis was conducted using Community Health Outcome and Health Determinants Database 1.3 from Korea Centers for Disease Control and Prevention. A total of 229 districts was included in the analysis. We compared the ordinary least squares (OLS) and the geographically weighted regression (GWR) to explore the associated factors of walking practice. MGWR 2.2.1 software was used to explore the spatial distribution of walking practice and modeling the GWR. Results: Walking practice had spatial variations across the country. The results showed that the GWR model had better accommodation of spatial autocorrelation than the OLS model. The GWR results indicated that different predictors of walking practice across regions of Korea. Conclusion: The findings of this study may provide insight to nursing researchers, health professionals, and policy makers in planning health programs to promote walking practices in their respective communities.

• Land and Housing Review
• /
• v.4 no.1
• /
• pp.33-42
• /
• 2013

This study which are recognized that the lack of empirical research about the efficiency of the elements of environmentally friendly housing development planned presented housing design elements and policies to revitalize for the reduction of greenhouse gas emissions by analyzing the effectiveness of reduction of greenhouse gas output. In addition, it used various models of DEA which are accepted until now effective technique to evaluate the performance of the organization. In conclusion, there are effective 5 regionals which are Seoul, Incheon, Ulsan, South Chungcheong Province, South Gyeongsang Province. other regionals was analyzed to be inefficient. The conclusion from this study are as follows: First, in case of 11 regionals which are analyzed to be inefficient, they have to difference plan elements to make up. So each region should establish strategy to complement vulnerability. Second, not only internal architectural factors but institutional, and external environmental factors also affect the reduction of greenhouse gas emissions. And weighted scores also were moderately high. But levels of weighted scores still less than the ratio of Good quality housing. So it can be determined that evaluation of individual architecture still considered important. It need to pay more attention to the operating system and the external environmental factors.

• MALSORI
• /
• no.44
• /
• pp.105-113
• /
• 2002

In Korean, each digit is monosyllable and some pairs are known to have high confusability, causing performance degradation of connected digit recognition systems. To improve the performance, in this paper, we employ various discriminant analyses (DA) including Linear DA (LDA), Weighted Pairwise Scatter LDA WPS-LDA), Heteroscedastic Discriminant Analysis (HDA), and Maximum Likelihood Linear Transformation (MLLT). We also examine several combinations of various DA for additional performance improvement. Experimental results show that applying any DA mentioned above improves the string accuracy, but the amount of improvement of each DA method varies according to the model complexity or number of mixtures per state. Especially, more than 20% of string error reduction is achieved by applying MLLT after WPS-LDA, compared with the baseline system, when class level of DA is defined as a tied state and 1 mixture per state is used.

• Journal of Environmental Impact Assessment
• /
• v.18 no.3
• /
• pp.143-150
• /
• 2009

The present study investigates the effects of the flight paths of military aircraft on noise map and its WECPNL(Weighted Equivalent Continuous Perceived Noise Level) distribution. Aircraft noise modeling and simulation have been performed on a Korean military air base by means of INM(Integrated Noise Model) with the input data of airfield location, aircraft specifications, flight paths and aircraft's operation schedules. The result of noise modelling has been verified in comparison with the result of measured noise level. The flight path of military aircraft, as the key parameter of the present study, was modeled by combining takeoff, overfly, approach and touch-and-go modes. The present INM simulations have been conducted for various flight path cases with different takeoff, approach modes and overfly modes. The simulation results showed that the change of flight path can remarkably affect the noise influence region and the WECPNL distribution around the airfield.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.6
• /
• pp.55-65
• /
• 1998

In industrial fields, RTP(rapid thermal processing) system is widely used for improving the oxidation and the annealing in semiconductor manufacturing process. The main control factors are temperature control of wafer and uniformity in the wafer. In this paper, we propose an $H^{\infty}$ controller design of RTP system satisfying robust stability and performance using weighted mixed sensitivity miniimization and loop shaping technique. And we need reduction technique because of the difficulty of implementation with the obtained high order controller for original model and reduced models, namely, Hankel, square-root balanced, and Schur balanced methods. An example is proposed to show the validity of the proposed method.

• Journal of the Korea Safety Management & Science
• /
• v.16 no.3
• /
• pp.451-469
• /
• 2014

The purpose of this study is to implement and develop the integrated Economic Value Added (EVA) and Time-Driven Activity-Based Costing (TDABC) model to seek both improvement of Net Operating Profit Less Adjusted Tax (NOPLAT) and reduction of Capital Charge (CC). Net Operating Profit Less Adjusted Tax (NOPLAT) can be maximized by reducing the indirect cost of an unused resource capacity increased by Cost Capacity Ratio (CCR) of TDABC. On the other hand, Capital Charge (CC) can be minimized by improving the efficiency of Invested Capital (IC) considered by Weighted Average Cost of Capital (WACC) of EVA. In addition, the integrated system of TDABC using Balance Scorecard (BSC) and EVA is developed by linking between the lagging indicators and the three leading indicators. The three leading indicators include customer, internal process and growth and learning perspectives whereas the lagging indicator includes NOPLAT and CC in terms of financial perspective. When the Critical Success Factor (CSF) of BSC is cascading as a cause and an effect relationship, time driver of TDABC and capital driver of EVA can be used efficiently as Key Performance Indicator (KPI) of BSC. For a better understanding of the proposed EVA/TDABC model and BSC/EVA/TDABC model, numerical examples are derived from this paper. From the proposed model, the time driver of TDABC and the capital driver of EVA are known to lessen indirect cost from comprehensive income statement when increasing the efficiency of operating IC from the statement of financial position with unified KPI cascading of aligned BSC CSFs.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.5
• /
• pp.533-539
• /
• 2019

This study proposes a malware classification model that can handle arbitrary length input data using the Microsoft Malware Classification Challenge dataset. We are based on imaging existing data from malware. The proposed model generates a lot of images when malware data is large, and generates a small image of small data. The generated image is learned as time series data by Dynamic RNN. The output value of the RNN is classified into malware by using only the highest weighted output by applying the Attention technique, and learning the RNN output value by Residual CNN again. Experiments on the proposed model showed a Micro-average F1 score of 92% in the validation data set. Experimental results show that the performance of a model capable of learning and classifying arbitrary length data can be verified without special feature extraction and dimension reduction.

• Structural Monitoring and Maintenance
• /
• v.5 no.2
• /
• pp.243-260
• /
• 2018

Detection of damages in fibre reinforced plastic (FRP) composite structures is important from the safety and serviceability point of view. Usually, damage is realized as a local reduction of stiffness and if dynamic responses of the structure are sensitive enough to such changes in stiffness, then a well posed inverse problem can provide an efficient solution to the damage detection problem. Usually, such inverse problems are solved within the framework of pattern recognition. Support Vector Machine (SVM) Algorithm is one such methodology, which minimizes the weighted differences between the experimentally observed dynamic responses and those computed using the finite element model- by optimizing appropriately chosen parameters, such as stiffness. A damage detection strategy is hereby proposed using SVM which perform stepwise by first locating and then determining the severity of the damage. The SVM algorithm uses simulations of only a limited number of damage scenarios and trains the algorithm in such a way so as to detect damages at unknown locations by recognizing the pattern of changes in dynamic responses. A rectangular fiber reinforced plastic composite plate has been investigated both numerically and experimentally to observe the efficiency of the SVM algorithm for damage detection. Experimentally determined modal responses, such as natural frequencies and mode shapes are used as observable parameters. The results are encouraging since a high percentage of damage cases have been successfully determined using the proposed algorithm.

• The Journal of Society for e-Business Studies
• /
• v.17 no.4
• /
• pp.141-153
• /
• 2012

According to the environment change of market, supply chain network needs to be redesigned for efficient provision of product within the budget constraint. Also, it is desired that the customer satisfaction such as on time delivery should be considered as an important element at redesigning of supply chain network in addition to the cost reduction. In this paper redesign of supply chain network for its expansion is treated as a problem situation and a related mathematical model is suggested. Moreover, the numerical examples about the total weighted distance of the redesigned supply chain network are presented with various budget constraints by using genetic algorithm to help the managerial decision.