Commenced in January 2007
Frequency: Monthly
Edition: International
Paper Count: 6259

Search results for: generalized extreme value distribution

6259 An Extension of the Generalized Extreme Value Distribution

Authors: Serge Provost, Abdous Saboor

Abstract:

A q-analogue of the generalized extreme value distribution which includes the Gumbel distribution is introduced. The additional parameter q allows for increased modeling flexibility. The resulting distribution can have a finite, semi-infinite or infinite support. It can also produce several types of hazard rate functions. The model parameters are determined by making use of the method of maximum likelihood. It will be shown that it compares favourably to three related distributions in connection with the modeling of a certain hydrological data set.

Keywords: extreme value theory, generalized extreme value distribution, goodness-of-fit statistics, Gumbel distribution

Procedia PDF Downloads 299
6258 Modeling of Maximum Rainfall Using Poisson-Generalized Pareto Distribution in Kigali, Rwanda

Authors: Emmanuel Iyamuremye

Abstract:

Extreme rainfall events have caused significant damage to agriculture, ecology, and infrastructure, disruption of human activities, injury, and loss of life. They also have significant social, economic, and environmental consequences because they considerably damage urban as well as rural areas. Early detection of extreme maximum rainfall helps to implement strategies and measures, before they occur, hence mitigating the consequences. Extreme value theory has been used widely in modeling extreme rainfall and in various disciplines, such as financial markets, the insurance industry, failure cases. Climatic extremes have been analyzed by using either generalized extreme value (GEV) or generalized Pareto (GP) distributions, which provides evidence of the importance of modeling extreme rainfall from different regions of the world. In this paper, we focused on Peak Over Thresholds approach, where the Poisson-generalized Pareto distribution is considered as the proper distribution for the study of the exceedances. This research also considers the use of the generalized Pareto (GP) distribution with a Poisson model for arrivals to describe peaks over a threshold. The research used statistical techniques to fit models that used to predict extreme rainfall in Kigali. The results indicate that the proposed Poisson-GP distribution provides a better fit to maximum monthly rainfall data. Further, the Poisson-GP models are able to estimate various return levels. The research also found a slow increase in return levels for maximum monthly rainfall for higher return periods, and further, the intervals are increasingly wider as the return period is increasing.

Keywords: exceedances, extreme value theory, generalized Pareto distribution, Poisson generalized Pareto distribution

Procedia PDF Downloads 94
6257 Generalized Extreme Value Regression with Binary Dependent Variable: An Application for Predicting Meteorological Drought Probabilities

Authors: Retius Chifurira

Abstract:

Logistic regression model is the most used regression model to predict meteorological drought probabilities. When the dependent variable is extreme, the logistic model fails to adequately capture drought probabilities. In order to adequately predict drought probabilities, we use the generalized linear model (GLM) with the quantile function of the generalized extreme value distribution (GEVD) as the link function. The method maximum likelihood estimation is used to estimate the parameters of the generalized extreme value (GEV) regression model. We compare the performance of the logistic and the GEV regression models in predicting drought probabilities for Zimbabwe. The performance of the regression models are assessed using the goodness-of-fit tests, namely; relative root mean square error (RRMSE) and relative mean absolute error (RMAE). Results show that the GEV regression model performs better than the logistic model, thereby providing a good alternative candidate for predicting drought probabilities. This paper provides the first application of GLM derived from extreme value theory to predict drought probabilities for a drought-prone country such as Zimbabwe.

Keywords: generalized extreme value distribution, general linear model, mean annual rainfall, meteorological drought probabilities

Procedia PDF Downloads 150
6256 Extreme Temperature Forecast in Mbonge, Cameroon Through Return Level Analysis of the Generalized Extreme Value (GEV) Distribution

Authors: Nkongho Ayuketang Arreyndip, Ebobenow Joseph

Abstract:

In this paper, temperature extremes are forecast by employing the block maxima method of the generalized extreme value (GEV) distribution to analyse temperature data from the Cameroon Development Corporation (CDC). By considering two sets of data (raw data and simulated data) and two (stationary and non-stationary) models of the GEV distribution, return levels analysis is carried out and it was found that in the stationary model, the return values are constant over time with the raw data, while in the simulated data the return values show an increasing trend with an upper bound. In the non-stationary model, the return levels of both the raw data and simulated data show an increasing trend with an upper bound. This clearly shows that although temperatures in the tropics show a sign of increase in the future, there is a maximum temperature at which there is no exceedance. The results of this paper are very vital in agricultural and environmental research.

Keywords: forecasting, generalized extreme value (GEV), meteorology, return level

Procedia PDF Downloads 414
6255 A Comparative Study of Generalized Autoregressive Conditional Heteroskedasticity (GARCH) and Extreme Value Theory (EVT) Model in Modeling Value-at-Risk (VaR)

Authors: Longqing Li

Abstract:

The paper addresses the inefficiency of the classical model in measuring the Value-at-Risk (VaR) using a normal distribution or a Student’s t distribution. Specifically, the paper focuses on the one day ahead Value-at-Risk (VaR) of major stock market’s daily returns in US, UK, China and Hong Kong in the most recent ten years under 95% confidence level. To improve the predictable power and search for the best performing model, the paper proposes using two leading alternatives, Extreme Value Theory (EVT) and a family of GARCH models, and compares the relative performance. The main contribution could be summarized in two aspects. First, the paper extends the GARCH family model by incorporating EGARCH and TGARCH to shed light on the difference between each in estimating one day ahead Value-at-Risk (VaR). Second, to account for the non-normality in the distribution of financial markets, the paper applies Generalized Error Distribution (GED), instead of the normal distribution, to govern the innovation term. A dynamic back-testing procedure is employed to assess the performance of each model, a family of GARCH and the conditional EVT. The conclusion is that Exponential GARCH yields the best estimate in out-of-sample one day ahead Value-at-Risk (VaR) forecasting. Moreover, the discrepancy of performance between the GARCH and the conditional EVT is indistinguishable.

Keywords: Value-at-Risk, Extreme Value Theory, conditional EVT, backtesting

Procedia PDF Downloads 286
6254 Extreme Value Modelling of Ghana Stock Exchange Indices

Authors: Kwabena Asare, Ezekiel N. N. Nortey, Felix O. Mettle

Abstract:

Modelling of extreme events has always been of interest in fields such as hydrology and meteorology. However, after the recent global financial crises, appropriate models for modelling of such rare events leading to these crises have become quite essential in the finance and risk management fields. This paper models the extreme values of the Ghana Stock Exchange All-Shares indices (2000-2010) by applying the Extreme Value Theory to fit a model to the tails of the daily stock returns data. A conditional approach of the EVT was preferred and hence an ARMA-GARCH model was fitted to the data to correct for the effects of autocorrelation and conditional heteroscedastic terms present in the returns series, before EVT method was applied. The Peak Over Threshold (POT) approach of the EVT, which fits a Generalized Pareto Distribution (GPD) model to excesses above a certain selected threshold, was employed. Maximum likelihood estimates of the model parameters were obtained and the model’s goodness of fit was assessed graphically using Q-Q, P-P and density plots. The findings indicate that the GPD provides an adequate fit to the data of excesses. The size of the extreme daily Ghanaian stock market movements were then computed using the Value at Risk (VaR) and Expected Shortfall (ES) risk measures at some high quantiles, based on the fitted GPD model.

Keywords: extreme value theory, expected shortfall, generalized pareto distribution, peak over threshold, value at risk

Procedia PDF Downloads 502
6253 Modelling Operational Risk Using Extreme Value Theory and Skew t-Copulas via Bayesian Inference

Authors: Betty Johanna Garzon Rozo, Jonathan Crook, Fernando Moreira

Abstract:

Operational risk losses are heavy tailed and are likely to be asymmetric and extremely dependent among business lines/event types. We propose a new methodology to assess, in a multivariate way, the asymmetry and extreme dependence between severity distributions, and to calculate the capital for Operational Risk. This methodology simultaneously uses (i) several parametric distributions and an alternative mix distribution (the Lognormal for the body of losses and the Generalized Pareto Distribution for the tail) via extreme value theory using SAS®, (ii) the multivariate skew t-copula applied for the first time for operational losses and (iii) Bayesian theory to estimate new n-dimensional skew t-copula models via Markov chain Monte Carlo (MCMC) simulation. This paper analyses a newly operational loss data set, SAS Global Operational Risk Data [SAS OpRisk], to model operational risk at international financial institutions. All the severity models are constructed in SAS® 9.2. We implement the procedure PROC SEVERITY and PROC NLMIXED. This paper focuses in describing this implementation.

Keywords: operational risk, loss distribution approach, extreme value theory, copulas

Procedia PDF Downloads 550
6252 Point Estimation for the Type II Generalized Logistic Distribution Based on Progressively Censored Data

Authors: Rana Rimawi, Ayman Baklizi

Abstract:

Skewed distributions are important models that are frequently used in applications. Generalized distributions form a class of skewed distributions and gain widespread use in applications because of their flexibility in data analysis. More specifically, the Generalized Logistic Distribution with its different types has received considerable attention recently. In this study, based on progressively type-II censored data, we will consider point estimation in type II Generalized Logistic Distribution (Type II GLD). We will develop several estimators for its unknown parameters, including maximum likelihood estimators (MLE), Bayes estimators and linear estimators (BLUE). The estimators will be compared using simulation based on the criteria of bias and Mean square error (MSE). An illustrative example of a real data set will be given.

Keywords: point estimation, type II generalized logistic distribution, progressive censoring, maximum likelihood estimation

Procedia PDF Downloads 156
6251 Frequency Analysis of Minimum Ecological Flow and Gage Height in Indus River Using Maximum Likelihood Estimation

Authors: Tasir Khan, Yejuan Wan, Kalim Ullah

Abstract:

Hydrological frequency analysis has been conducted to estimate the minimum flow elevation of the Indus River in Pakistan to protect the ecosystem. The Maximum likelihood estimation (MLE) technique is used to estimate the best-fitted distribution for Minimum Ecological Flows at nine stations of the Indus River in Pakistan. The four selected distributions, Generalized Extreme Value (GEV) distribution, Generalized Logistics (GLO) distribution, Generalized Pareto (GPA) distribution, and Pearson type 3 (PE3) are fitted in all sites, usually used in hydro frequency analysis. Compare the performance of these distributions by using the goodness of fit tests, such as the Kolmogorov Smirnov test, Anderson darling test, and chi-square test. The study concludes that the Maximum Likelihood Estimation (MLE) method recommended that GEV and GPA are the most suitable distributions which can be effectively applied to all the proposed sites. The quantiles are estimated for the return periods from 5 to 1000 years by using MLE, estimations methods. The MLE is the robust method for larger sample sizes. The results of these analyses can be used for water resources research, including water quality management, designing irrigation systems, determining downstream flow requirements for hydropower, and the impact of long-term drought on the country's aquatic system.

Keywords: minimum ecological flow, frequency distribution, indus river, maximum likelihood estimation

Procedia PDF Downloads 41
6250 Evaluation of Best-Fit Probability Distribution for Prediction of Extreme Hydrologic Phenomena

Authors: Karim Hamidi Machekposhti, Hossein Sedghi

Abstract:

The probability distributions are the best method for forecasting of extreme hydrologic phenomena such as rainfall and flood flows. In this research, in order to determine suitable probability distribution for estimating of annual extreme rainfall and flood flows (discharge) series with different return periods, precipitation with 40 and discharge with 58 years time period had been collected from Karkheh River at Iran. After homogeneity and adequacy tests, data have been analyzed by Stormwater Management and Design Aid (SMADA) software and residual sum of squares (R.S.S). The best probability distribution was Log Pearson Type III with R.S.S value (145.91) and value (13.67) for peak discharge and Log Pearson Type III with R.S.S values (141.08) and (8.95) for maximum discharge in Jelogir Majin and Pole Zal stations, respectively. The best distribution for maximum precipitation in Jelogir Majin and Pole Zal stations was Log Pearson Type III distribution with R.S.S values (1.74&1.90) and then Pearson Type III distribution with R.S.S values (1.53&1.69). Overall, the Log Pearson Type III distributions are acceptable distribution types for representing statistics of extreme hydrologic phenomena in Karkheh River at Iran with the Pearson Type III distribution as a potential alternative.

Keywords: Karkheh River, Log Pearson Type III, probability distribution, residual sum of squares

Procedia PDF Downloads 160
6249 VaR or TCE: Explaining the Preferences of Regulators

Authors: Silvia Faroni, Olivier Le Courtois, Krzysztof Ostaszewski

Abstract:

While a lot of research concentrates on the merits of VaR and TCE, which are the two most classic risk indicators used by financial institutions, little has been written on explaining why regulators favor the choice of VaR or TCE in their set of rules. In this paper, we investigate the preferences of regulators with the aim of understanding why, for instance, a VaR with a given confidence level is ultimately retained. Further, this paper provides equivalence rules that explain how a given choice of VaR can be equivalent to a given choice of TCE. Then, we introduce a new risk indicator that extends TCE by providing a more versatile weighting of the constituents of probability distribution tails. All of our results are illustrated using the generalized Pareto distribution.

Keywords: generalized pareto distribution, generalized tail conditional expectation, regulator preferences, risk measure

Procedia PDF Downloads 120
6248 The Generalized Pareto Distribution as a Model for Sequential Order Statistics

Authors: Mahdy ‎Esmailian, Mahdi ‎Doostparast, Ahmad ‎Parsian

Abstract:

‎In this article‎, ‎sequential order statistics (SOS) censoring type II samples coming from the generalized Pareto distribution are considered‎. ‎Maximum likelihood (ML) estimators of the unknown parameters are derived on the basis of the available multiple SOS data‎. ‎Necessary conditions for existence and uniqueness of the derived ML estimates are given‎. Due to complexity in the proposed likelihood function‎, ‎a useful re-parametrization is suggested‎. ‎For illustrative purposes‎, ‎a Monte Carlo simulation study is conducted and an illustrative example is analysed‎.

Keywords: bayesian estimation‎, generalized pareto distribution‎, ‎maximum likelihood estimation‎, sequential order statistics

Procedia PDF Downloads 463
6247 Nonstationary Modeling of Extreme Precipitation in the Wei River Basin, China

Authors: Yiyuan Tao

Abstract:

Under the impact of global warming together with the intensification of human activities, the hydrological regimes may be altered, and the traditional stationary assumption was no longer satisfied. However, most of the current design standards of water infrastructures were still based on the hypothesis of stationarity, which may inevitably result in severe biases. Many critical impacts of climate on ecosystems, society, and the economy are controlled by extreme events rather than mean values. Therefore, it is of great significance to identify the non-stationarity of precipitation extremes and model the precipitation extremes in a nonstationary framework. The Wei River Basin (WRB), located in a continental monsoon climate zone in China, is selected as a case study in this study. Six extreme precipitation indices were employed to investigate the changing patterns and stationarity of precipitation extremes in the WRB. To identify if precipitation extremes are stationary, the Mann-Kendall trend test and the Pettitt test, which is used to examine the occurrence of abrupt changes are adopted in this study. Extreme precipitation indices series are fitted with non-stationary distributions that selected from six widely used distribution functions: Gumbel, lognormal, Weibull, gamma, generalized gamma and exponential distributions by means of the time-varying moments model generalized additive models for location, scale and shape (GAMLSS), where the distribution parameters are defined as a function of time. The results indicate that: (1) the trends were not significant for the whole WRB, but significant positive/negative trends were still observed in some stations, abrupt changes for consecutive wet days (CWD) mainly occurred in 1985, and the assumption of stationarity is invalid for some stations; (2) for these nonstationary extreme precipitation indices series with significant positive/negative trends, the GAMLSS models are able to capture well the temporal variations of the indices, and perform better than the stationary model. Finally, the differences between the quantiles of nonstationary and stationary models are analyzed, which highlight the importance of nonstationary modeling of precipitation extremes in the WRB.

Keywords: extreme precipitation, GAMLSSS, non-stationary, Wei River Basin

Procedia PDF Downloads 85
6246 A Bayesian Model with Improved Prior in Extreme Value Problems

Authors: Eva L. Sanjuán, Jacinto Martín, M. Isabel Parra, Mario M. Pizarro

Abstract:

In Extreme Value Theory, inference estimation for the parameters of the distribution is made employing a small part of the observation values. When block maxima values are taken, many data are discarded. We developed a new Bayesian inference model to seize all the information provided by the data, introducing informative priors and using the relations between baseline and limit parameters. Firstly, we studied the accuracy of the new model for three baseline distributions that lead to a Gumbel extreme distribution: Exponential, Normal and Gumbel. Secondly, we considered mixtures of Normal variables, to simulate practical situations when data do not adjust to pure distributions, because of perturbations (noise).

Keywords: bayesian inference, extreme value theory, Gumbel distribution, highly informative prior

Procedia PDF Downloads 155
6245 Effect of Outliers in Assessing Significant Wave Heights Through a Time-Dependent GEV Model

Authors: F. Calderón-Vega, A. D. García-Soto, C. Mösso

Abstract:

Recorded significant wave heights sometimes exhibit large uncommon values (outliers) that can be associated with extreme phenomena such as hurricanes and cold fronts. In this study, some extremely large wave heights recorded in NOAA buoys (National Data Buoy Center, noaa.gov) are used to investigate their effect in the prediction of future wave heights associated with given return periods. Extreme waves are predicted through a time-dependent model based on the so-called generalized extreme value distribution. It is found that the outliers do affect the estimated wave heights. It is concluded that a detailed inspection of outliers is envisaged to determine whether they are real recorded values since this will impact defining design wave heights for coastal protection purposes.

Keywords: GEV model, non-stationary, seasonality, outliers

Procedia PDF Downloads 151
6244 Frequency Analysis Using Multiple Parameter Probability Distributions for Rainfall to Determine Suitable Probability Distribution in Pakistan

Authors: Tasir Khan, Yejuan Wang

Abstract:

The study of extreme rainfall events is very important for flood management in river basins and the design of water conservancy infrastructure. Evaluation of quantiles of annual maximum rainfall (AMRF) is required in different environmental fields, agriculture operations, renewable energy sources, climatology, and the design of different structures. Therefore, the annual maximum rainfall (AMRF) was performed at different stations in Pakistan. Multiple probability distributions, log normal (LN), generalized extreme value (GEV), Gumbel (max), and Pearson type3 (P3) were used to find out the most appropriate distributions in different stations. The L moments method was used to evaluate the distribution parameters. Anderson darling test, Kolmogorov- Smirnov test, and chi-square test showed that two distributions, namely GUM (max) and LN, were the best appropriate distributions. The quantile estimate of a multi-parameter PD offers extreme rainfall through a specific location and is therefore important for decision-makers and planners who design and construct different structures. This result provides an indication of these multi-parameter distribution consequences for the study of sites and peak flow prediction and the design of hydrological maps. Therefore, this discovery can support hydraulic structure and flood management.

Keywords: RAMSE, multiple frequency analysis, annual maximum rainfall, L-moments

Procedia PDF Downloads 42
6243 Calibration of Hybrid Model and Arbitrage-Free Implied Volatility Surface

Authors: Kun Huang

Abstract:

This paper investigates whether the combination of local and stochastic volatility models can be calibrated exactly to any arbitrage-free implied volatility surface of European option. The risk neutral Brownian Bridge density is applied for calibration of the leverage function of our Hybrid model. Furthermore, the tails of marginal risk neutral density are generated by Generalized Extreme Value distribution in order to capture the properties of asset returns. The local volatility is generated from the arbitrage-free implied volatility surface using stochastic volatility inspired parameterization.

Keywords: arbitrage free implied volatility, calibration, extreme value distribution, hybrid model, local volatility, risk-neutral density, stochastic volatility

Procedia PDF Downloads 229
6242 Parameter Estimation for the Mixture of Generalized Gamma Model

Authors: Wikanda Phaphan

Abstract:

Mixture generalized gamma distribution is a combination of two distributions: generalized gamma distribution and length biased generalized gamma distribution. These two distributions were presented by Suksaengrakcharoen and Bodhisuwan in 2014. The findings showed that probability density function (pdf) had fairly complexities, so it made problems in estimating parameters. The problem occurred in parameter estimation was that we were unable to calculate estimators in the form of critical expression. Thus, we will use numerical estimation to find the estimators. In this study, we presented a new method of the parameter estimation by using the expectation – maximization algorithm (EM), the conjugate gradient method, and the quasi-Newton method. The data was generated by acceptance-rejection method which is used for estimating α, β, λ and p. λ is the scale parameter, p is the weight parameter, α and β are the shape parameters. We will use Monte Carlo technique to find the estimator's performance. Determining the size of sample equals 10, 30, 100; the simulations were repeated 20 times in each case. We evaluated the effectiveness of the estimators which was introduced by considering values of the mean squared errors and the bias. The findings revealed that the EM-algorithm had proximity to the actual values determined. Also, the maximum likelihood estimators via the conjugate gradient and the quasi-Newton method are less precision than the maximum likelihood estimators via the EM-algorithm.

Keywords: conjugate gradient method, quasi-Newton method, EM-algorithm, generalized gamma distribution, length biased generalized gamma distribution, maximum likelihood method

Procedia PDF Downloads 184
6241 Population Size Estimation Based on the GPD

Authors: O. Anan, D. Böhning, A. Maruotti

Abstract:

The purpose of the study is to estimate the elusive target population size under a truncated count model that accounts for heterogeneity. The purposed estimator is based on the generalized Poisson distribution (GPD), which extends the Poisson distribution by adding a dispersion parameter. Thus, it becomes an useful model for capture-recapture data where concurrent events are not homogeneous. In addition, it can account for over-dispersion and under-dispersion. The ratios of neighboring frequency counts are used as a tool for investigating the validity of whether generalized Poisson or Poisson distribution. Since capture-recapture approaches do not provide the zero counts, the estimated parameters can be achieved by modifying the EM-algorithm technique for the zero-truncated generalized Poisson distribution. The properties and the comparative performance of proposed estimator were investigated through simulation studies. Furthermore, some empirical examples are represented insights on the behavior of the estimators.

Keywords: capture, recapture methods, ratio plot, heterogeneous population, zero-truncated count

Procedia PDF Downloads 402
6240 New Hybrid Method to Model Extreme Rainfalls

Authors: Youness Laaroussi, Zine Elabidine Guennoun, Amine Amar

Abstract:

Modeling and forecasting dynamics of rainfall occurrences constitute one of the major topics, which have been largely treated by statisticians, hydrologists, climatologists and many other groups of scientists. In the same issue, we propose in the present paper a new hybrid method, which combines Extreme Values and fractal theories. We illustrate the use of our methodology for transformed Emberger Index series, constructed basing on data recorded in Oujda (Morocco). The index is treated at first by Peaks Over Threshold (POT) approach, to identify excess observations over an optimal threshold u. In the second step, we consider the resulting excess as a fractal object included in one dimensional space of time. We identify fractal dimension by the box counting. We discuss the prospect descriptions of rainfall data sets under Generalized Pareto Distribution, assured by Extreme Values Theory (EVT). We show that, despite of the appropriateness of return periods given by POT approach, the introduction of fractal dimension provides accurate interpretation results, which can ameliorate apprehension of rainfall occurrences.

Keywords: extreme values theory, fractals dimensions, peaks Over threshold, rainfall occurrences

Procedia PDF Downloads 325
6239 The Normal-Generalized Hyperbolic Secant Distribution: Properties and Applications

Authors: Hazem M. Al-Mofleh

Abstract:

In this paper, a new four-parameter univariate continuous distribution called the Normal-Generalized Hyperbolic Secant Distribution (NGHS) is defined and studied. Some general and structural distributional properties are investigated and discussed, including: central and non-central n-th moments and incomplete moments, quantile and generating functions, hazard function, Rényi and Shannon entropies, shapes: skewed right, skewed left, and symmetric, modality regions: unimodal and bimodal, maximum likelihood (MLE) estimators for the parameters. Finally, two real data sets are used to demonstrate empirically its flexibility and prove the strength of the new distribution.

Keywords: bimodality, estimation, hazard function, moments, Shannon’s entropy

Procedia PDF Downloads 299
6238 Joint Probability Distribution of Extreme Water Level with Rainfall and Temperature: Trend Analysis of Potential Impacts of Climate Change

Authors: Ali Razmi, Saeed Golian

Abstract:

Climate change is known to have the potential to impact adversely hydrologic patterns for variables such as rainfall, maximum and minimum temperature and sea level rise. Long-term average of these climate variables could possibly change over time due to climate change impacts. In this study, trend analysis was performed on rainfall, maximum and minimum temperature and water level data of a coastal area in Manhattan, New York City, Central Park and Battery Park stations to investigate if there is a significant change in the data mean. Partial Man-Kendall test was used for trend analysis. Frequency analysis was then performed on data using common probability distribution functions such as Generalized Extreme Value (GEV), normal, log-normal and log-Pearson. Goodness of fit tests such as Kolmogorov-Smirnov are used to determine the most appropriate distributions. In flood frequency analysis, rainfall and water level data are often separately investigated. However, in determining flood zones, simultaneous consideration of rainfall and water level in frequency analysis could have considerable effect on floodplain delineation (flood extent and depth). The present study aims to perform flood frequency analysis considering joint probability distribution for rainfall and storm surge. First, correlation between the considered variables was investigated. Joint probability distribution of extreme water level and temperature was also investigated to examine how global warming could affect sea level flooding impacts. Copula functions were fitted to data and joint probability of water level with rainfall and temperature for different recurrence intervals of 2, 5, 25, 50, 100, 200, 500, 600 and 1000 was determined and compared with the severity of individual events. Results for trend analysis showed increase in long-term average of data that could be attributed to climate change impacts. GEV distribution was found as the most appropriate function to be fitted to the extreme climate variables. The results for joint probability distribution analysis confirmed the necessity for incorporation of both rainfall and water level data in flood frequency analysis.

Keywords: climate change, climate variables, copula, joint probability

Procedia PDF Downloads 313
6237 Analysis of the Statistical Characterization of Significant Wave Data Exceedances for Designing Offshore Structures

Authors: Rui Teixeira, Alan O’Connor, Maria Nogal

Abstract:

The statistical theory of extreme events is progressively a topic of growing interest in all the fields of science and engineering. The changes currently experienced by the world, economic and environmental, emphasized the importance of dealing with extreme occurrences with improved accuracy. When it comes to the design of offshore structures, particularly offshore wind turbines, the importance of efficiently characterizing extreme events is of major relevance. Extreme events are commonly characterized by extreme values theory. As an alternative, the accurate modeling of the tails of statistical distributions and the characterization of the low occurrence events can be achieved with the application of the Peak-Over-Threshold (POT) methodology. The POT methodology allows for a more refined fit of the statistical distribution by truncating the data with a minimum value of a predefined threshold u. For mathematically approximating the tail of the empirical statistical distribution the Generalised Pareto is widely used. Although, in the case of the exceedances of significant wave data (H_s) the 2 parameters Weibull and the Exponential distribution, which is a specific case of the Generalised Pareto distribution, are frequently used as an alternative. The Generalized Pareto, despite the existence of practical cases where it is applied, is not completely recognized as the adequate solution to model exceedances over a certain threshold u. References that set the Generalised Pareto distribution as a secondary solution in the case of significant wave data can be identified in the literature. In this framework, the current study intends to tackle the discussion of the application of statistical models to characterize exceedances of wave data. Comparison of the application of the Generalised Pareto, the 2 parameters Weibull and the Exponential distribution are presented for different values of the threshold u. Real wave data obtained in four buoys along the Irish coast was used in the comparative analysis. Results show that the application of the statistical distributions to characterize significant wave data needs to be addressed carefully and in each particular case one of the statistical models mentioned fits better the data than the others. Depending on the value of the threshold u different results are obtained. Other variables of the fit, as the number of points and the estimation of the model parameters, are analyzed and the respective conclusions were drawn. Some guidelines on the application of the POT method are presented. Modeling the tail of the distributions shows to be, for the present case, a highly non-linear task and, due to its growing importance, should be addressed carefully for an efficient estimation of very low occurrence events.

Keywords: extreme events, offshore structures, peak-over-threshold, significant wave data

Procedia PDF Downloads 212
6236 Classical and Bayesian Inference of the Generalized Log-Logistic Distribution with Applications to Survival Data

Authors: Abdisalam Hassan Muse, Samuel Mwalili, Oscar Ngesa

Abstract:

A generalized log-logistic distribution with variable shapes of the hazard rate was introduced and studied, extending the log-logistic distribution by adding an extra parameter to the classical distribution, leading to greater flexibility in analysing and modeling various data types. The proposed distribution has a large number of well-known lifetime special sub-models such as; Weibull, log-logistic, exponential, and Burr XII distributions. Its basic mathematical and statistical properties were derived. The method of maximum likelihood was adopted for estimating the unknown parameters of the proposed distribution, and a Monte Carlo simulation study is carried out to assess the behavior of the estimators. The importance of this distribution is that its tendency to model both monotone (increasing and decreasing) and non-monotone (unimodal and bathtub shape) or reversed “bathtub” shape hazard rate functions which are quite common in survival and reliability data analysis. Furthermore, the flexibility and usefulness of the proposed distribution are illustrated in a real-life data set and compared to its sub-models; Weibull, log-logistic, and BurrXII distributions and other parametric survival distributions with 3-parmaeters; like the exponentiated Weibull distribution, the 3-parameter lognormal distribution, the 3- parameter gamma distribution, the 3-parameter Weibull distribution, and the 3-parameter log-logistic (also known as shifted log-logistic) distribution. The proposed distribution provided a better fit than all of the competitive distributions based on the goodness-of-fit tests, the log-likelihood, and information criterion values. Finally, Bayesian analysis and performance of Gibbs sampling for the data set are also carried out.

Keywords: hazard rate function, log-logistic distribution, maximum likelihood estimation, generalized log-logistic distribution, survival data, Monte Carlo simulation

Procedia PDF Downloads 153
6235 Regional Flood-Duration-Frequency Models for Norway

Authors: Danielle M. Barna, Kolbjørn Engeland, Thordis Thorarinsdottir, Chong-Yu Xu

Abstract:

Design flood values give estimates of flood magnitude within a given return period and are essential to making adaptive decisions around land use planning, infrastructure design, and disaster mitigation. Often design flood values are needed at locations with insufficient data. Additionally, in hydrologic applications where flood retention is important (e.g., floodplain management and reservoir design), design flood values are required at different flood durations. A statistical approach to this problem is a development of a regression model for extremes where some of the parameters are dependent on flood duration in addition to being covariate-dependent. In hydrology, this is called a regional flood-duration-frequency (regional-QDF) model. Typically, the underlying statistical distribution is chosen to be the Generalized Extreme Value (GEV) distribution. However, as the support of the GEV distribution depends on both its parameters and the range of the data, special care must be taken with the development of the regional model. In particular, we find that the GEV is problematic when developing a GAMLSS-type analysis due to the difficulty of proposing a link function that is independent of the unknown parameters and the observed data. We discuss these challenges in the context of developing a regional QDF model for Norway.

Keywords: design flood values, bayesian statistics, regression modeling of extremes, extreme value analysis, GEV

Procedia PDF Downloads 34
6234 Pressure Distribution, Load Capacity, and Thermal Effect with Generalized Maxwell Model in Journal Bearing Lubrication

Authors: M. Guemmadi, A. Ouibrahim

Abstract:

This numerical investigation aims to evaluate how a viscoelastic lubricant described by a generalized Maxwell model, affects the pressure distribution, the load capacity and thermal effect in a journal bearing lubrication. We use for the purpose the CFD package software completed by adapted user define functions (UDFs) to solve the coupled equations of momentum, of energy and of the viscoelastic model (generalized Maxwell model). Two parameters, viscosity and relaxation time are involved to show how viscoelasticity substantially affect the pressure distribution, the load capacity and the thermal transfer by comparison to Newtonian lubricant. These results were also compared with the available published results.

Keywords: journal bearing, lubrication, Maxwell model, viscoelastic fluids, computational modelling, load capacity

Procedia PDF Downloads 503
6233 A Bivariate Inverse Generalized Exponential Distribution and Its Applications in Dependent Competing Risks Model

Authors: Fatemah A. Alqallaf, Debasis Kundu

Abstract:

The aim of this paper is to introduce a bivariate inverse generalized exponential distribution which has a singular component. The proposed bivariate distribution can be used when the marginals have heavy-tailed distributions, and they have non-monotone hazard functions. Due to the presence of the singular component, it can be used quite effectively when there are ties in the data. Since it has four parameters, it is a very flexible bivariate distribution, and it can be used quite effectively for analyzing various bivariate data sets. Several dependency properties and dependency measures have been obtained. The maximum likelihood estimators cannot be obtained in closed form, and it involves solving a four-dimensional optimization problem. To avoid that, we have proposed to use an EM algorithm, and it involves solving only one non-linear equation at each `E'-step. Hence, the implementation of the proposed EM algorithm is very straight forward in practice. Extensive simulation experiments and the analysis of one data set have been performed. We have observed that the proposed bivariate inverse generalized exponential distribution can be used for modeling dependent competing risks data. One data set has been analyzed to show the effectiveness of the proposed model.

Keywords: Block and Basu bivariate distributions, competing risks, EM algorithm, Marshall-Olkin bivariate exponential distribution, maximum likelihood estimators

Procedia PDF Downloads 100
6232 Estimating The Population Mean by Using Stratified Double Extreme Ranked Set Sample

Authors: Mahmoud I. Syam, Kamarulzaman Ibrahim, Amer I. Al-Omari

Abstract:

Stratified double extreme ranked set sampling (SDERSS) method is introduced and considered for estimating the population mean. The SDERSS is compared with the simple random sampling (SRS), stratified ranked set sampling (SRSS) and stratified simple set sampling (SSRS). It is shown that the SDERSS estimator is an unbiased of the population mean and more efficient than the estimators using SRS, SRSS and SSRS when the underlying distribution of the variable of interest is symmetric or asymmetric.

Keywords: double extreme ranked set sampling, extreme ranked set sampling, ranked set sampling, stratified double extreme ranked set sampling

Procedia PDF Downloads 419
6231 Parameter Estimation of Gumbel Distribution with Maximum-Likelihood Based on Broyden Fletcher Goldfarb Shanno Quasi-Newton

Authors: Dewi Retno Sari Saputro, Purnami Widyaningsih, Hendrika Handayani

Abstract:

Extreme data on an observation can occur due to unusual circumstances in the observation. The data can provide important information that can’t be provided by other data so that its existence needs to be further investigated. The method for obtaining extreme data is one of them using maxima block method. The distribution of extreme data sets taken with the maxima block method is called the distribution of extreme values. Distribution of extreme values is Gumbel distribution with two parameters. The parameter estimation of Gumbel distribution with maximum likelihood method (ML) is difficult to determine its exact value so that it is necessary to solve the approach. The purpose of this study was to determine the parameter estimation of Gumbel distribution with quasi-Newton BFGS method. The quasi-Newton BFGS method is a numerical method used for nonlinear function optimization without constraint so that the method can be used for parameter estimation from Gumbel distribution whose distribution function is in the form of exponential doubel function. The quasi-New BFGS method is a development of the Newton method. The Newton method uses the second derivative to calculate the parameter value changes on each iteration. Newton's method is then modified with the addition of a step length to provide a guarantee of convergence when the second derivative requires complex calculations. In the quasi-Newton BFGS method, Newton's method is modified by updating both derivatives on each iteration. The parameter estimation of the Gumbel distribution by a numerical approach using the quasi-Newton BFGS method is done by calculating the parameter values that make the distribution function maximum. In this method, we need gradient vector and hessian matrix. This research is a theory research and application by studying several journals and textbooks. The results of this study obtained the quasi-Newton BFGS algorithm and estimation of Gumbel distribution parameters. The estimation method is then applied to daily rainfall data in Purworejo District to estimate the distribution parameters. This indicates that the high rainfall that occurred in Purworejo District decreased its intensity and the range of rainfall that occurred decreased.

Keywords: parameter estimation, Gumbel distribution, maximum likelihood, broyden fletcher goldfarb shanno (BFGS)quasi newton

Procedia PDF Downloads 285
6230 Parametric Modeling for Survival Data with Competing Risks Using the Generalized Gompertz Distribution

Authors: Noora Al-Shanfari, M. Mazharul Islam

Abstract:

The cumulative incidence function (CIF) is a fundamental approach for analyzing survival data in the presence of competing risks, which estimates the marginal probability for each competing event. Parametric modeling of CIF has the advantage of fitting various shapes of CIF and estimates the impact of covariates with maximum efficiency. To calculate the total CIF's covariate influence using a parametric model., it is essential to parametrize the baseline of the CIF. As the CIF is an improper function by nature, it is necessary to utilize an improper distribution when applying parametric models. The Gompertz distribution, which is an improper distribution, is limited in its applicability as it only accounts for monotone hazard shapes. The generalized Gompertz distribution, however, can adapt to a wider range of hazard shapes, including unimodal, bathtub, and monotonic increasing or decreasing hazard shapes. In this paper, the generalized Gompertz distribution is used to parametrize the baseline of the CIF, and the parameters of the proposed model are estimated using the maximum likelihood approach. The proposed model is compared with the existing Gompertz model using the Akaike information criterion. Appropriate statistical test procedures and model-fitting criteria will be used to test the adequacy of the model. Both models are applied to the ‘colon’ dataset, which is available in the “biostat3” package in R.

Keywords: competing risks, cumulative incidence function, improper distribution, parametric modeling, survival analysis

Procedia PDF Downloads 48