• 한국음향학회지
• /
• 제18권4호
• /
• pp.47-52
• /
• 1999

본 논문에서는 특징 파라미터의 분산과 인식성능에 대한 기여도를 고려하여 각 특징 파라미터를 가중시키는 방법을 제안하였다. 각 특징 파라미터의 인식률에 비례하게 전체 기여도를 설정하고, 각 특징 파라미터의 분산에 따라 가중요인을 설정하였다. 전체 기여도와 분산에 따른 가중요인을 사용하여 각 특징 파라미터의 상태별 가중치를 설정하였다. 제안한 방법의 유효성을 살펴보기 위해 유사음소 단위의 HMM 음성인식시스템을 사용하여 인식실험을 하였다. 인식실험에서 제안한 방법으로 가중치를 설정하였을 경우에 인식률이 7.7% 향상됨을 볼 수 있었다.

• Communications for Statistical Applications and Methods
• /
• 제29권4호
• /
• pp.413-420
• /
• 2022

In this study we adapt discrete weibull regression model for clustered count data. Discrete weibull regression model has an attractive feature that it can handle both under and over dispersion data. We analyzed the eighth Korean National Health and Nutrition Examination Survey (KNHANES VIII) from 2019 to assess the factors influencing the 1 month outpatient stay in 17 different regions. We compared the results using clustered discrete Weibull regression model with those of Poisson, negative binomial, generalized Poisson and Conway-maxwell Poisson regression models, which are widely used in count data analyses. The results show that the clustered discrete Weibull regression model using random intercept model gives the best fit. Simulation study is also held to investigate the performance of the clustered discrete weibull model under various dispersion setting and zero inflated probabilities. In this paper it is shown that using a random effect with discrete Weibull regression can flexibly model count data with various dispersion without the risk of making wrong assumptions about the data dispersion.

• Genomics & Informatics
• /
• 제17권4호
• /
• pp.39.1-39.20
• /
• 2019

Analyzing patterns in data points embedded in linear and non-linear feature spaces is considered as one of the common research problems among different research areas, for example: data mining, machine learning, pattern recognition, and multivariate analysis. In this paper, data points are heterogeneous sets of biosequences (composite data points). A composite data point is a set of ordinary data points (e.g., set of feature vectors). We theoretically extend the derivation of the largest generalized eigenvalue-based distance metric D_ij(γ₁) in any linear and non-linear feature spaces. We prove that D_ij(γ₁) is a metric under any linear and non-linear feature transformation function. We show the sufficiency and efficiency of using the decision rule $\bar{{\delta}}_{{\Xi}i}$(i.e., mean of D_ij(γ₁)) in classification of heterogeneous sets of biosequences compared with the decision rules min_𝚵iand median_𝚵i. We analyze the impact of linear and non-linear transformation functions on classifying/clustering collections of heterogeneous sets of biosequences. The impact of the length of a sequence in a heterogeneous sequence-set generated by simulation on the classification and clustering results in linear and non-linear feature spaces is empirically shown in this paper. We propose a new concept: the limiting dispersion map of the existing clusters in heterogeneous sets of biosequences embedded in linear and nonlinear feature spaces, which is based on the limiting distribution of nucleotide compositions estimated from real data sets. Finally, the empirical conclusions and the scientific evidences are deduced from the experiments to support the theoretical side stated in this paper.

• Smart Structures and Systems
• /
• 제16권1호
• /
• pp.213-222
• /
• 2015

Early detection and precise location of leakage is of great importance for life-cycle maintenance and management of municipal pipeline system. In the past few years, acoustic emission (AE) techniques have demonstrated to be an excellent tool for on-line leakage detection. Regarding the multi-mode and frequency dispersion characteristics of AE signals propagating along a pipeline, the direct cross-correlation technique that assumes the constant AE propagation velocity does not perform well in practice for acoustic leak location. This paper presents an improved cross-correlation method based on wavelet transform, with due consideration of the frequency dispersion characteristics of AE wave and the contribution of different mode. Laboratory experiments conducted to simulate pipeline gas leakage and investigate the frequency spectrum signatures of AE leak signals. By comparing with the other methods for leak location identification, the feasibility and superiority of the proposed method are verified.

• 대한전기학회논문지:전기물성ㆍ응용부문C
• /
• 제50권3호
• /
• pp.121-129
• /
• 2001

In this paper, a new efficient feature extraction method based on Daubechies discrete wavelet transform is presented. This paper especially deals with the assessment of process statistical parameter using the features extracted from the wavelet coefficients of measured acoustic emission signals. Since the parameter assessment using all wavelet coefficients will often turn out leads to inefficient or inaccurate results, we selected that level-3 stage of multi decomposition in discrete wavelet transform. We make use of the feature extraction parameter namely, maximum value of acoustic emission signal, average value, dispersion, skewness, kurtosis, etc. The effectiveness of this new method has been verified on ability a diagnosis transformer go through feature extraction in stage of aging(the early period, the middle period, the last period)

• Asian Journal of Atmospheric Environment
• /
• 제10권3호
• /
• pp.125-136
• /
• 2016

In this study, the effects of the mesh barrier on the free dispersion of ammonia were numerically investigated under different atmospheric conditions. This study presents the detail and flow feature of the dispersion of ammonia through the mesh barrier on various free stream conditions to decline and limit the toxic danger of the ammonia. It is assumed that the dispersion of the ammonia occurred through the leakage in the pipeline. Parametric studies were conducted on the performance of the mesh barrier by using the Reynolds-averaged Navier-Stokes equations with realizable k-${\varepsilon}$ turbulence model. Numerical simulations of ammonia dispersion in the presence of mesh barrier revealed significant results in a fully turbulent free stream condition. The results clearly show that the flow behavior was found to be a direct result of mesh size and ammonia dispersion is highly influenced by these changes in flow patterns in downstream. In fact, the flow regime becomes laminar as flow passes through mesh barrier. According to the results, the mesh barrier decreased the maximum concentration of the ammonia gas and limited the risk zone (more than 500 ppm) lower than 2 m height. Furthermore, a significant reduction occurs in the slope of the upper boundary of $NH_3$ risk zone distribution at downstream when a mesh barrier is presented. Thus, this device highly restricts the leak distribution of ammonia in the industrial plan.

• 한국대기환경학회지
• /
• 제23권4호
• /
• pp.405-419
• /
• 2007

A series of tracer experiments for the evaluation of atmospheric dispersion was performed over the urban area of Seoul using two inert, non-deposition perfluorocarbon (PMCH and m-PDCH) gases during three years campaign on 2002, 2003 and 2005. 30 sampling sites for collecting these tracers were located along two arcs of 2.5 and 5 kilometers downwind from the release point. About ten measurements which each lasted for 2 hours or 4 hours were made over the two consecutive days during each campaign. CALPUFF and MM5 meteorological model were applied to evaluate the urban dispersion in detail. Size of Modeling domain was $27\;km{\times}23\;km$ and the fine nest in the modeling domain had a grid size of 0.5 km. The results showed that CALPUFF dispersion model had a tendency to estimate tracer concentrations about $2{\sim}5$ times less than those of ambient samples under many conditions. These consistent inaccuracy in urban dispersion was attributed to inherent inaccuracy and lack of details in terrain data at urban area.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.9-9
• /
• 2000

Accurate counting and packaging pills is one of the most fundamental works of the pharmaceutical industry. But it is so labor consuming and very hard to be automated. As the pharmaceutical industry is growing bigger, the need of counting and packaging automation is increasing to obtain effective mass production. Precise and quick sensing is required in the counting and processing of quickly dropping pills to improve the productivity. There are many trials for this automation and automatic machine. But the performance of the existing counting machine varies with the size, shape and the dispersion degree of pills In this research, authors design the counting and packing machine of medicinal pills that is more accurate and highly trustworthy After getting analog signal from optical sensor, pill passage is discriminated from chosen characteristic feature using microprocessor.

• 한국항해항만학회지
• /
• 제28권3호
• /
• pp.185-191
• /
• 2004

평균법과 클러스터링은 다속성 평가문제에서 널리 쓰이고 있는 중요한 데이터 마이닝 기법들이다. 그러나, 다양한 다속성 평가 문제에서 데이터 마이닝을 할 때, 데이터들의 특징은 그 중요성이 달라질 수 있기 때문에 이러한 데이터의 중요도 차이를 고려해야 할 필요가 있다. 따라서, 이러한 기법들은 데이터의 선택 및 중요도 등과 같이 그 특징을 얼마나 잘 반영하는 지가 중요하다. 게다가, 산술평균법의 경우에는 우선순위 및 가중치로 정의되는 평가구조에서 적합한 결과를 산출하기에는 한계가 있을뿐 만 아니라, 평가자 그룹별 특징을 반영하기 곤란하다. 따라서, 본 연구에서는 기하학적 도형을 바탕으로 유사도를 평가하여, 평가자 그룹별로 특징지어지는 이산적인 환경에서의 평균을 산출하는 알고리즘을 제안하였다. 본 알고리즘의 핵심사항 중 하나는, 항목별 우선순위의 혼돈없이 유사도를 평가할 수 있다는 점이다.

• Asian Journal of Atmospheric Environment
• /
• 제8권2호
• /
• pp.81-88
• /
• 2014

Noise-barriers on both sides of the roadway (hereafter referred to as double noise-barriers), are a common feature along roads in Korea, and these are expected to have important effects on the near-road air pollution dispersion of vehicle emissions. This study evaluated the double noise-barrier impact on near-road air pollution dispersion, using a FLUENT computational fluid dynamics (CFD) model. The realizable k-${\varepsilon}$ model in FLUENT CFD code was used to simulate vehicle air pollutant dispersion, in around 11 cases of double noise-barriers. The simulated concentration profiles and surface concentrations under no barrier cases were compared with the experimental results. The results of the simulated flows show the following three regimes in this study: isolated roughness (H/W=0.05), wake interface (H/W=0.1), and skimming flow (H/W>0.15). The results also show that the normalized average concentrations at surface (z=1 m) between the barriers increase with increasing double noise-barrier height; however, normalized average concentrations at the top position between the barriers decrease with increasing barrier height. It was found that the double noise-barrier decreases normalized average concentrations of leeward positions, ranging from 0.8 (H/W=0.1, wake interface) to 0.1 (H/W=0.5, skimming flow) times lower than that of the no barrier case, at 10 x/h downwind position; and ranging from 1.0 (H/W=0.1) to 0.4 (H/W=0.5) times lower than that of the no barrier case, at 60 x/h downwind position.