• Structural Engineering and Mechanics
• /
• v.74 no.1
• /
• pp.55-67
• /
• 2020

The common practice to predict the characteristic structural load effects (LEs) in long reference periods is to employ the extreme value theory (EVT) for building limit distributions. However, most applications ignore that LEs are driven by multiple loading events and thus do not have the identical distribution, a prerequisite for EVT. In this study, we propose the composite extreme value modeling approach using clustering to (a) cluster initial blended samples into finite identical distributed subsamples using the finite mixture model, expectation-maximization algorithm, and the Akaike information criterion; (b) combine limit distributions of subsamples into a composite prediction equation using the generalized Pareto distribution based on a joint threshold. The proposed approach was validated both through numerical examples with known solutions and engineering applications of bridge traffic LEs on a long-span bridge. The results indicate that a joint threshold largely benefits the composite extreme value modeling, many appropriate tail approaching models can be used, and the equation form is simply the sum of the weighted models. In numerical examples, the proposed approach using clustering generated accurate extrema prediction of any reference period compared with the known solutions, whereas the common practice of employing EVT without clustering on the mixture data showed large deviations. Real-world bridge traffic LEs are driven by multi-events and present multipeak distributions, and the proposed approach is more capable of capturing the tendency of tailed LEs than the conventional approach. The proposed approach is expected to have wide applications to general problems such as samples that are driven by multiple events and that do not have the identical distribution.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.39 no.1
• /
• pp.19-26
• /
• 2015

This paper proposes a design method for improving the frequency-domain performance reliability of dynamic systems with uncertain and degrading components. Discrete frequencies are used in this method as surrogates for the frequency band of interest, and the conformance of the frequency responses to the specification at these frequencies is utilized to model the frequency-domain performance reliability. A meta-model for the frequency responses, an extreme-value event, and the set-theory are integrated to improve the computational efficiency of the reliability estimation. In addition, a sample-based approach is presented to evaluate and optimize the estimated performance reliability. A case study of a vibration absorber system showed that the proposed design method has engineering applications.

• Wind and Structures
• /
• v.25 no.3
• /
• pp.261-282
• /
• 2017

The probabilistic information of directional extreme wind speeds is important for precisely estimating the design wind loads on structures. A new joint probability distribution model of directional extreme wind speeds is established based on observed wind-speed data using multivariate extreme value theory with the t-Copula function in the present study. At first, the theoretical deficiencies of the Gaussian-Copula and Gumbel-Copula models proposed by previous researchers for the joint probability distribution of directional extreme wind speeds are analysed. Then, the t-Copula model is adopted to solve this deficiency. Next, these three types of Copula models are discussed and evaluated with Spearman's rho, the parametric bootstrap test and the selection criteria based on the empirical Copula. Finally, the extreme wind speeds for a given return period are predicted by the t-Copula model with observed wind-speed records from several areas and the influence of dependence among directional extreme wind speeds on the predicted results is discussed.

• Knowledge Management Research
• /
• v.22 no.2
• /
• pp.247-268
• /
• 2021

Investors mainly use PER and PBR among financial ratios for valuation and investment decision-making. I conduct an analysis of two basic financial ratios from a statistical perspective. Financial ratios contain key accounting numbers which reflect firm fundamentals and are useful for valuation or risk analysis such as enterprise credit evaluation and default prediction. The distribution of financial data tends to be extremely heavy-tailed, and PER and PBR show exceedingly high level of kurtosis and their extreme cases often contain significant information on financial risk. In this respect, Extreme Value Theory is required to fit its right tail more precisely. I introduce not only GPD but exGPD. GPD is conventionally preferred model in Extreme Value Theory and exGPD is log-transformed distribution of GPD. exGPD has recently proposed as an alternative of GPD(Lee and Kim, 2019). First, I conduct a simulation for comparing performances of the two distributions using the goodness of fit measures and the estimation of 90-99% percentiles. I also conduct an empirical analysis of Information Technology firms in Korea. Finally, exGPD shows better performance especially for PBR, suggesting that exGPD could be an alternative for GPD for the analysis of financial ratios.

• International Journal of Highway Engineering
• /
• v.18 no.1
• /
• pp.1-11
• /
• 2016

PURPOSES : This study estimated the load effect of a single heavy truck to develop a live load model for the design and assessment of bridges located on an expressway with a limited truck entry weight. METHODS : The statistical estimation methods for the live load effect acting on a bridge by a heavy vehicle are reviewed, and applications using the actual measurement data for trucks traveling on an expressway are presented. The weight estimation of a single vehicle and its effect on a bridge are fundamental elements in the construction of a live load model. Two statistical estimation methods for the application of extrapolation in a probabilistic study and an additional estimation method that adopts the extreme value theory are reviewed. RESULTS : The proposed methods are applied to the traffic data measured on an expressway. All of the estimation methods yield similar results using the data measured when the weight limit has been relatively well observed because of the rigid enforcement of the weight regulation. On the other hand, when the estimations are made using overweight traffic data, the resulting values differ with the estimation method. CONCLUSIONS : The estimation methods based on the extreme distribution theory and the modified procedure presented in this paper can yield reasonable values for the maximum weight of a single truck, which can be applied in both the design and evaluation of a bridge on an expressway.

• The Korean Journal of Applied Statistics
• /
• v.27 no.4
• /
• pp.655-665
• /
• 2014

An adequate understanding and response to natural hazards such as heat wave, heavy rainfall and severe drought is required. We apply extreme value theory to analyze these abnormal weather phenomena. It is common for extremes in climatic data to be nonstationary in space and time. In this paper, we analyze summer rainfall data in South Korea using exceedance values over thresholds estimated by quantile regression with location information and time as covariates. We group weather stations in South Korea into 5 clusters and t extreme value models to threshold exceedances for each cluster under the assumption of independence in space and time as well as estimates of uncertainty for spatial dependence as proposed in Northrop and Jonathan (2011).

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.36 no.2
• /
• pp.223-230
• /
• 2012

The strength of brittle material has commonly been characterized by a normal distribution or Weibull distribution, but it may fit the gamma distribution for some material. The use of an extreme value distribution is proper when the largest values of a set of stresses dominate the failure of the material. This paper presents a formula for reliability estimation based on stress-strength interference theory that is applicable when the strength of material is distributed like a gamma distribution and the stress is distributed like an extreme value distribution. We verified the validity of the equation for the reliability estimation by examining the relationships among the factor of safety, the coefficient of variation, and the reliability. The required minimum factor of safety and the highest allowable coefficient of variation of stress can be estimated by choosing an objective reliability and estimating the reliabilities obtained for various factors of safety and coefficients of variation.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2014.10a
• /
• pp.467-471
• /
• 2014

The natural frequencies of a mechanical system are determined by the system parameters such as masses and stiffness of the system. Since material irregularities and manufacturing tolerances always exist in most of practical engineering situations, the system parameters always have uncertainties. As the uncertainties of the parameters increase, the uncertainties of the system natural frequencies also increases. Then, the reliability of the system deteriorates. So, the uncertainty of the system natural frequencies should be analyzed accurately and considered in the design of the system. In order to analyze the uncertainty of the system natural frequencies employing most of existing uncertainty analysis methods, the probability distributions of the uncertain system parameters should be identified. In most practical situations, however, identification of the probability distributions is almost impossible because of limited time and cost. For that case, the reliability should be estimated based on finite samples of the system parameters. In this paper, sample based reliability estimation method employing extreme value theory was proposed. Using the proposed estimation method, sample based reliability design of the system natural frequencies was conducted.

• Journal of the Korean Statistical Society
• /
• v.32 no.4
• /
• pp.313-335
• /
• 2003

Hourly ozone data are available for 73 stations in South Korea from January, 1988 to August, 1998. We are interested in detecting trends in both the mean levels and the extremes of ozone, and in determining how these trends vary over the country. The latter aspect means that we also have to understand the spatial dependence of ozone. In this connection, therefore, we examine in this paper the following features: determining trends in mean ozone levels at individual stations and combination across stations; determining trends in extreme ozone levels at individual stations and combination across stations; spatial modeling of trends in mean and extreme ozone levels.

• Communications for Statistical Applications and Methods
• /
• v.26 no.3
• /
• pp.239-259
• /
• 2019

The transmuted generalized extreme value (TGEV) distribution was first introduced by Aryal and Tsokos (Nonlinear Analysis: Theory, Methods & Applications, 71, 401-407, 2009) and applied by Nascimento et al. (Hacettepe Journal of Mathematics and Statistics, 45, 1847-1864, 2016). However, they did not give explicit expressions for all the moments, tail behaviour, quantiles, survival and risk functions and order statistics. The TGEV distribution is a more flexible model than the simple GEV distribution to model extreme or rare events because the right tail of the TGEV is heavier than the GEV. In addition the TGEV distribution can adjusted various forms of asymmetry. In this article, explicit expressions for these measures of the TGEV are obtained. The tail behavior and the survival and risk functions were determined for positive gamma, the moments for nonzero gamma and the moment generating function for zero gamma. The performance of the maximum likelihood estimators (MLEs) of the TGEV parameters were tested through a series of Monte Carlo simulation experiments. In addition, the model was used to fit three real data sets related to financial returns.