• International Journal of Control, Automation, and Systems
• /
• 제6권1호
• /
• pp.109-118
• /
• 2008

We present two estimators for discrete non-Gaussian and nonstationary probability density estimation based on a dynamic Bayesian network (DBN). The first estimator is for off line computation and consists of a DBN whose transition distribution is represented in terms of kernel functions. The estimator parameters are the weights and shifts of the kernel functions. The parameters are determined through a recursive learning algorithm using maximum likelihood (ML) estimation. The second estimator is a DBN whose parameters form the transition probabilities. We use an asymptotically convergent, recursive, on-line algorithm to update the parameters using observation data. The DBN calculates the state probabilities using the estimated parameters. We provide examples that demonstrate the usefulness and simplicity of the two proposed estimators.

• 한국정보처리학회논문지
• /
• 제6권3호
• /
• pp.766-772
• /
• 1999

With the increase of portable system and high-density IC, power consumption of VLSI circuits is very important factor in design process. Power estimation is required in order to estimate the power consumption. A simple and correct solution of power estimation is to use circuit simulation. But it is very time consuming and inefficient way. Probabilistic method has been proposed to overcome this problem. Transition density using probability was an efficient method to estimate power consumption using BDD and Boolean difference. But it is difficult to build the BDD and compute complex Boolean difference. In this paper, we proposed Propowest. Propowest is building a digraph of circuit, and easy and fast in computing transition density by using modified COP algorithm. Propowest provides an efficient way for power estimation.

• 대한전기학회논문지
• /
• 제45권2호
• /
• pp.325-328
• /
• 1996

In this paper, we proposed an estimation method of a posterior probability and PDF(Probability density function) using a feed forward neural network and code books of VQ(vector quantization). In this study, We estimates a posterior probability and probability density function, which compose a new parameter with well-known Mel cepstrum and verificate the performance for the five vowels taking from syllables by NN(neural network) and PNN(probabilistic neural network). In case of new parameter, showed the best result by probabilistic neural network and recognition rates are average 83.02%.

• 한국구조물진단유지관리공학회 논문집
• /
• 제15권2호
• /
• pp.70-78
• /
• 2011

본 논문은 변형률 계측데이터를 사용하는 신뢰성 및 정확성을 증진된 BWIM(Bridge Weigh-In-Motion) 알고리즘을 개발하고, 이를 교량에 대한 다양한 실험을 통해 검증하고자 하는 것이다. 본 논문에서는 밀도추정함수와 평균보정계수를 이용한 BWIM 알고리즘을 제시한다. 밀도추정함수는 다축하중을 추정할 때 신뢰할 수 있게 적용할 수 있음을 입증하였으며, 평균보정계수는 이론 계산된 모멘트와 계측된 변형률에서 계산한 모멘트 사이의 전반적인 오차를 최소화하기 위해 적용된다. 개발된 알고리즘은 수치예제, 실내모형실험 그리고 다주형 합성교량에 대한 현장실험을 통해 성공적으로 검증하였다.

• Journal of the Korean Statistical Society
• /
• 제28권4호
• /
• pp.443-455
• /
• 1999

A simulation-based approach to estimating the probability of an arbitrary region under a multivariate normal distribution is developed. In specific, the probability is expressed as the ratio of the unrestricted and the restricted multivariate normal density functions, where the restriction is given by the region whose probability is of interest. The density function of the restricted distribution is then estimated by using a sample generated from the Gibbs sampling algorithm.

• Journal of the Korean Statistical Society
• /
• 제27권1호
• /
• pp.91-100
• /
• 1998

A method of estimating probability density using regression tools is presented here. It is based on equal-length binning and locally weighted approximate likelihood for bin counts. The method is particularly useful for densities with bounded supports, where it automatically corrects edge effects without using boundary kernels.

• 한국초전도ㆍ저온공학회논문지
• /
• 제12권3호
• /
• pp.16-20
• /
• 2010

For large scale power applications of HTS conductor, it is getting more important to have a stacked HTS coated conductor with low loss and large current capacity. But it was not easy to measure some electric properties. Stabilizer free YBCO CC for striated/ stacked conductors is easily burned out during the measurement of the critical current density because it has no stabilizer and it is difficult to set-up the current lead and voltage taps because it has many pieces of YBCO CC in a conductor. Instead of direct measuring the critical current of a stacked HTS coated conductor, indirect estimation from measuring a magnetization loss of HTS coated conductor could be useful for practical estimation of the critical current. The magnetization loss of a superconductor is supposed to be affected by a full penetrating magnetic field, and it tends to show an inflection point at the full penetrating magnetic field when we generate the graph of magnetization loss vs. external magnetic field. The full penetrating magnetic field depends on the shape of the conductor and its critical current density, so we can estimate the effective critical current density from measuring the magnetization loss. In this paper, to prove the effectiveness of this indirect estimation of the critical current, we prepared several different kinds of YBCO CC(coated conductor) including a stacked conductor short samples and measured the magnetization losses and the critical currents of each sample by using linked pick up coils and direct voltage measurement with transport current respectively.

• 한국환경보건학회지
• /
• 제34권2호
• /
• pp.131-136
• /
• 2008

The variation of the particle size distribution and density as well as the chemical composition of aerosols is important to evaluate the particles. This study measured and analyzed airborne particles using a scanning mobility particle sizer (SMPS) system and an aerodynamic particle sizer (APS) at the University of Seoul during every season. The highest particle number concentration of airborne particles less than $0.9\;{\mu}m$, occurred in winter, while the highest particle number concentration of airborne particles more than $0.9\;{\mu}m$, occurred in spring. Mass concentration appeared highest at spring. Also, when we compared $\beta$-ray's mass concentration with calculated mass concentration by using the SMPS-APS system during each season, density of the winter is $1.92\;g/cm^3$, spring density is $1.64\;g/cm^3$, fall density is $1.57\;g/cm^3$. We found out that PM10 density was differ every season. However, while the calculated density is whole density for PM10 the density of each diameter was different. In this study the density estimation equation of the QCM cascade impactor measured mass concentration of each diameter.

• 정보처리학회논문지A
• /
• 제9A권1호
• /
• pp.75-82
• /
• 2002

전력 소모 예측에 필요한 신호 전이 밀도를 구하기 위하여, 제로 지연 모델에 대한 견실한 신호 전이 밀도 전파 방법이 제시된다. 제로 기연 모델을 위한 전력 예측은 전력 소모의 하한 경계값을 위한 적절한 기준이다. 입력 특성이 일반적으로 설계 단계에 알려져 있지 않기 때문에 광범위한 입력 특성에 대한 견실한 예측은 전력 소모에 대하여 매우 중요하다. 본 연구에서는 기존의 신호 전이 예측 방법에 대하여 입력 및 출력의 변이 특성을 분석하고 이러한 분석 결과에 근거하여 새로운 견실한 신호 전이 밀도 전파 방법을 제안한다. 실제 회로에 적용하기 위하여 전력 예측의 정확성에 크게 영향을 미치는 재수렴성 경로를 고려한 알고리즘을 제안 및 연구한다. 실험에 의하면 제안한 방법이 기존의 방법과 비교할 때 더욱 양호한 견실성 및 종래의 방식에 상응하는 정확성과 경과 시간을 보여준다.

• Journal of Information Processing Systems
• /
• 제8권4호
• /
• pp.669-684
• /
• 2012

Discrete wavelet transforms are extensively preferred in biomedical signal processing for denoising, feature extraction, and compression. This paper presents a new denoising method based on the modeling of discrete wavelet coefficients of ECG in selected sub-bands with Kernel density estimation. The modeling provides a statistical distribution of information and noise. A Gaussian kernel with bounded support is used for modeling sub-band coefficients and thresholds and is estimated by placing a sliding window on a normalized cumulative density function. We evaluated this approach on offline noisy ECG records from the Cardiovascular Research Centre of the University of Glasgow and on records from the MIT-BIH Arrythmia database. Results show that our proposed technique has a more reliable physical basis and provides improvement in the Signal-to-Noise Ratio (SNR) and Percentage RMS Difference (PRD). The morphological information of ECG signals is found to be unaffected after employing denoising. This is quantified by calculating the mean square error between the feature vectors of original and denoised signal. MSE values are less than 0.05 for most of the cases.