• 한국환경과학회지
• /
• 제22권8호
• /
• pp.965-977
• /
• 2013

Reliable long-term streamflow forecasting is invaluable for water resource planning and management which allocates water supply according to the demand of water users. It is necessary to get probabilistic forecasts to establish risk-based reservoir operation policies. Probabilistic forecasts may be useful for the users who assess and manage risks according to decision-making responding forecasting results. Probabilistic forecasting of seasonal inflow to Andong dam is performed and assessed using selected predictors from sea surface temperature and 500 hPa geopotential height data. Categorical probability forecast by Piechota's method and logistic regression analysis, and probability forecast by conditional probability density function are used to forecast seasonal inflow. Kernel density function is used in categorical probability forecast by Piechota's method and probability forecast by conditional probability density function. The results of categorical probability forecasts are assessed by Brier skill score. The assessment reveals that the categorical probability forecasts are better than the reference forecasts. The results of forecasts using conditional probability density function are assessed by qualitative approach and transformed categorical probability forecasts. The assessment of the forecasts which are transformed to categorical probability forecasts shows that the results of the forecasts by conditional probability density function are much better than those of the forecasts by Piechota's method and logistic regression analysis except for winter season data.

• Journal of the Korean Statistical Society
• /
• 제13권2호
• /
• pp.87-106
• /
• 1984

It ${X_t}$ is a sequence of independent identically distributed normal random variables, then the conditional probability density of $X_1, X_2, \cdots, X_n$ given the first p+1 sample autocovariances converges to the maximum entropy probability density satisfying the corresponding covariance constraints as the length of the sample sequence tends to infinity. This establishes that the maximum entropy probability density and the associated Gaussian autoregressive process arise naturally as the answers of conditional limit problems.

• Journal of the Korean Data and Information Science Society
• /
• 제27권5호
• /
• pp.1155-1168
• /
• 2016

기상 자료의 경우 한 지역의 기후가 인접지역의 기후와 비슷한 양상을 띄고 각 지역의 확률 밀도 함수 (probability density function)가 잘 알려진 확률 모형을 따르지 않는다는 것이 알려져 있다. 본 논문에서는 이러한 특성을 고려하여 이상 기후 현상이 뚜렷히 나타나는 여름철 평균 극한 기온(extreme temperature)의 확률 밀도 함수를 추정하고자 한다. 이를 위하여 공간적 상관관계 (spatial correlation)를 고려하는 비모수 베이지안 (nonparametric Bayesian) 모형인 조건부 자기회귀 종추출 혼합모형 (mixtures of conditional autoregression species sampling model)을 이용하였다. 자료는 이스트앵글리아 대학교 (University of East Anglia)에서 제공하는 전 지구의 최대 기온과 최소 기온자료 중 우리나라에 해당하는 지역의 자료를 사용하였다.

• 한국지능시스템학회논문지
• /
• 제14권7호
• /
• pp.847-851
• /
• 2004

Bishop과 Nabnck에 의해 소개된 기존치 혼합 밀도 네트워크(Mixture Density Network)에서는 조건부 확률밀도 함수의 매개변수들(parameters)이 하나의 MLP(multi-layer perceptron)의 출력 벡터로 주어진다. 최근에는 변형된 혼합 밀도 네트워크(Modified Mixture Density Network)라고 하는 이름으로 조건부 확률밀도 함수의 선분포(priors), 조건부 평균(conditional means), 그리고 공분산(covariances) 등이 각각 독립적인 MLP의 출력벡터로 주어지는 경우를 다룬 연구가 보고된 바 있다. 본 논문에서는 조건부 평균이 입력에 관해 선형인 경우를 위한 버전에 대한 이론과 매트랩 프로그램 개발을 다룬다. 본 논문에서는 우선 일반적인 혼합 밀도 네트워크에 대해 간단히 설명하고, 혼합 밀도 네트워크의 출력인 다층 퍼셉트론의 매개변수를 각각 다른 다층 퍼셉트론에서 학습시키는 변형된 혼합 밀도 네트워크를 설명한 후, 각각 다른 다층 퍼셉트론을 통해 매개변수를 얻는 것은 동일하나 평균값은 선형함수를 통해 얻는 혼합 밀도 네트워크 버전을 소개한다. 그리고, 모의실험을 통하여 이러한 혼합 밀도 네트워크의 적용가능성에 대해 알아본다.

• Earthquakes and Structures
• /
• 제22권2호
• /
• pp.157-168
• /
• 2022

A moment-independent importance measure analysis approach was introduced to quantify the effects of structural uncertainty parameters on probabilistic seismic demands of simply supported girder bridges. Based on the probability distributions of main uncertainty parameters in bridges, conditional and unconditional bridge samples were constructed with Monte-Carlo sampling and analyzed in the OpenSees platform with a series of real seismic ground motion records. Conditional and unconditional probability density functions were developed using kernel density estimation with the results of nonlinear time history analysis of the bridge samples. Moment-independent importance measures of these uncertainty parameters were derived by numerical integrations with the conditional and unconditional probability density functions, and the uncertainty parameters were ranked in descending order of their importance. Different from Tornado diagram approach, the impacts of uncertainty parameters on the whole probability distributions of bridge seismic demands and the interactions of uncertainty parameters were considered simultaneously in the importance measure analysis approach. Results show that the interaction of uncertainty parameters had significant impacts on the seismic demand of components, and in some cases, it changed the most significant parameters for piers, bearings and abutments.

• 한국지능시스템학회논문지
• /
• 제19권3호
• /
• pp.311-316
• /
• 2009

본 논문에서는 파티클 필터(Particle Filter)를 사용한 모바일 감시 로봇을 위한 실시간 움직임 추정 알고리즘을 제안한다. 파티클 필터는 몬테카를로(Monte Carlo) 샘플링 방법을 기반으로 사전분포확률(Prior distribution probability)와 사후분포확률(Posterior distribution probability)을 가지는 베이지안 조건 확률 모델(Bayesian conditional probabilities model)을 사용하는 방법이다. 그러나 대부분의 파티클 필터에서는 초기 확률밀도(Prior probability density)를 임의로 정의하여 사용하지만, 본 논문에서는 Sum of Absolute Difference (SAD)를 이용하여 초기 확률밀도를 구하고, 이를 파티클 필터에 적용하여 모바일 감시 로봇 환경에서 임의로 움직이는 물체를 강인하게 실시간으로 추정하고 추적하는 시스템을 구현하였다.

• Structural Engineering and Mechanics
• /
• 제32권1호
• /
• pp.55-69
• /
• 2009

A quasi ideal importance sampling simulation method combined in the conditional expectation is proposed for the structural reliability estimation. The quasi ideal importance sampling joint probability density function (p.d.f.) is so composed on the basis of the ideal importance sampling concept as to be proportional to the conditional failure probability multiplied by the p.d.f. of the sampling variables. The respective marginal p.d.f.s of the ideal importance sampling joint p.d.f. are determined numerically by the simulations and partly by the piecewise integrations. The quasi ideal importance sampling simulations combined in the conditional expectation are executed to estimate the failure probabilities of structures with multiple failure surfaces and it is shown that the proposed method gives accurate estimations efficiently.

• Structural Engineering and Mechanics
• /
• 제57권1호
• /
• pp.21-43
• /
• 2016

A conditional probability based approach known as Particle Filter Method (PFM) is a powerful tool for system parameter identification. In this paper, PFM has been applied to identify the vehicle parameters based on response statistics of the bridge. The flexibility of vehicle model has been considered in the formulation of bridge-vehicle interaction dynamics. The random unevenness of bridge has been idealized as non homogeneous random process in space. The simulated response has been contaminated with artificial noise to reflect the field condition. The performance of the identification system has been examined for various measurement location, vehicle velocity, bridge surface roughness factor, noise level and assumption of prior probability density. Identified vehicle parameters are found reasonably accurate and reconstructed interactive force time history with identified parameters closely matches with the simulated results. The study also reveals that crude assumption of prior probability density function does not end up with an incorrect estimate of parameters except requiring longer time for the iterative process to converge.

• Genomics & Informatics
• /
• 제20권2호
• /
• pp.17.1-17.11
• /
• 2022

Genetic associations have been quantified using a number of statistical measures. Entropy-based mutual information may be one of the more direct ways of estimating the association, in the sense that it does not depend on the parametrization. For this purpose, both the entropy and conditional entropy of the phenotype distribution should be obtained. Quantitative traits, however, do not usually allow an exact evaluation of entropy. The estimation of entropy needs a probability density function, which can be approximated by kernel density estimation. We have investigated the proper sequence of procedures for combining the kernel density estimation and entropy estimation with a probability density function in order to calculate mutual information. Genotypes and their interactions were constructed to set the conditions for conditional entropy. Extensive simulation data created using three types of generating functions were analyzed using two different kernels as well as two types of multifactor dimensionality reduction and another probability density approximation method called m-spacing. The statistical power in terms of correct detection rates was compared. Using kernels was found to be most useful when the trait distributions were more complex than simple normal or gamma distributions. A full-scale genomic dataset was explored to identify associations using the 2-h oral glucose tolerance test results and γ-glutamyl transpeptidase levels as phenotypes. Clearly distinguishable single-nucleotide polymorphisms (SNPs) and interacting SNP pairs associated with these phenotypes were found and listed with empirical p-values.

• 충청수학회지
• /
• 제27권2호
• /
• pp.243-247
• /
• 2014

Let {$X_n$, $n{\geq}1$} be a sequence of i.i.d. random variables with absolutely continuous cumulative distribution function(cdf) F(x) and the corresponding probability density function(pdf) f(x). In this paper, we give characterizations of Pareto and Weibull distribution by considering conditional expectations of record values.