Improved Transferability of Data‐Driven Damage Models Through Sample Selection Bias Correction

Damage models for natural hazards are used for decision making on reducing and transferring risk. The damage estimates from these models depend on many variables and their complex sometimes nonlinear relationships with the damage. In recent years, data‐driven modeling techniques have been used to capture those relationships. The available data to build such models are often limited. Therefore, in practice it is usually necessary to transfer models to a different context. In this article, we show that this implies the samples used to build the model are often not fully representative for the situation where they need to be applied on, which leads to a “sample selection bias.” In this article, we enhance data‐driven damage models by applying methods, not previously applied to damage modeling, to correct for this bias before the machine learning (ML) models are trained. We demonstrate this with case studies on flooding in Europe, and typhoon wind damage in the Philippines. Two sample selection bias correction methods from the ML literature are applied and one of these methods is also adjusted to our problem. These three methods are combined with stochastic generation of synthetic damage data. We demonstrate that for both case studies, the sample selection bias correction techniques reduce model errors, especially for the mean bias error this reduction can be larger than 30%. The novel combination with stochastic data generation seems to enhance these techniques. This shows that sample selection bias correction methods are beneficial for damage model transfer.     

Book Reviews

Books reviewed:
M Hollander and D Wolfe, Nonparametric Statistical Methods
T Leonard and J.S.J Hsu, Bayesian Methods     

Quadratic estimators of covariance components in a multivariate mixed linear model

It is known that the Henderson Method III (Biometrics 9:226–252, 1953) is of special interest for the mixed linear models because the estimators of the variance components are unaffected by the parameters of the fixed factor (or factors). This article deals with generalizations and minor extensions of the results obtained for the univariate linear models. A MANOVA mixed model is presented in a convenient form and the covariance components estimators are given on finite dimensional linear spaces. The results use both the usual parametric representations and the coordinate-free approach of Kruskal (Ann Math Statist 39:70–75, 1968) and Eaton (Ann Math Statist 41:528–538, 1970). The normal equations are generalized and it is given a necessary and sufficient condition for the existence of quadratic unbiased estimators for covariance components in the considered model.     

The analysis of incomplete data using stochastic covariates

In the development of many diseases there are often associated random variables which continuously reflect the progress of a subject towards the final expression of the disease (failure). At any given time these processes, which we call stochastic covariates, may provide information about the current hazard and the remaining time to failure. Likewise, in situations when the specific times of key prior events are not known, such as the time of onset of an occult tumour or the time of infection with HIV-1, it may be possible to identify a stochastic covariate which reveals, indirectly, when the event of interest occurred. The analysis of carcinogenicity trials which involve occult tumours is usually based on the time of death or sacrifice and an indicator of tumour presence for each animal in the experiment. However, the size of an occult tumour observed at the endpoint represents data concerning tumour development which may convey additional information concerning both the tumour incidence rate and the rate of death to which tumour-bearing animals are subject. We develop a stochastic model for tumour growth and suggest different ways in which the effect of this growth on the hazard of failure might be modelled. Using a combined model for tumour growth and additive competing risks of death, we show that if this tumour size information is used, assumptions concerning tumour lethality, the context of observation or multiple sacrifice times are no longer necessary in order to estimate the tumour incidence rate. Parametric estimation based on the method of maximum likelihood is outlined and is applied to simulated data from the combined model. The results of this limited study confirm that use of the stochastic covariate tumour size results in more precise estimation of the incidence rate for occult tumours.     

均值绝对偏差资产组合选择模型的算法

对均值绝对偏差模型进行了简化,并利用一种旋转算法求解。这种算法比单纯形算法的计算简便,且计算量更小。利用上海和深圳股市1 072支股票70期周末收盘价所作的实验结果表明,对于资产无上界限制的模型,计算20个不同最优投资组合需要1 274次旋转运算,上界为10%时需要1 570次旋转运算,每 次旋转运算约需1 14171次加法和乘法运算。     

Physiologically Based Liver Modeling and Risk Assessment

Because of the inherent complexity of biological systems, there is often a choice between a number of apparently equally applicable physiologically based models to describe uptake and metabolism processes in toxicology or risk assessment. These models may fit the particular data sets of interest equally well, but may give quite different parameter estimates or predictions under different (extrapolated) conditions. Such competing models can be discriminated by a number of methods, including potential refutation by means of strategic experiments, and their ability to suitably incorporate all relevant physiological processes. For illustration, three currently used models for steady-state hepatic elimination--the venous equilibration model, the parallel tube model, and the distributed sinusoidal perfusion model--are reviewed and compared with particular reference to their application in the area of risk assessment. The ability of each of the models to describe and incorporate such physiological processes as protein binding, precursor-metabolite relations and hepatic zones of elimination, capillary recruitment, capillary heterogeneity, and intrahepatic shunting is discussed. Differences between the models in hepatic parameter estimation, extrapolation to different conditions, and interspecies scaling are discussed, and criteria for choosing one model over the others are presented. In this case, the distributed model provides the most general framework for describing physiological processes taking place in the liver, and has so far not been experimentally refuted, as have the other two models. These simpler models may, however, provide useful bounds on parameter estimates and on extrapolations and risk assessments.     

Ranking Nuclear and Radiological Terrorism Scenarios: The Italian Case

A quantitative criterion for ranking the different scenarios of nuclear and radiological terrorism has been developed. The aim of the model is not to predict terroristic events but only to indicate which scenario has the higher utility from the point of view of a terroristic organization in terms of balance between factors favoring and discouraging the attack, respectively. All these factors were quantified according to a scoring system that takes into account the logarithmic relationship between perceptions and stimuli. The criterion was applied to several scenarios, each of which was modeled in a simple but not trivial way in order to estimate the expected damage in terms of probable life losses from both radiative and nonradiative effects. The outcome from the ranking method indicates that the attractive scenario appears to be the detonation of a low yield improvised nuclear device in the metropolitan area of a major city.     

Conditional Likelihood Estimators for Hidden Markov Models and Stochastic Volatility Models

ABSTRACT.  This paper develops a new contrast process for parametric inference of general hidden Markov models, when the hidden chain has a non-compact state space. This contrast is based on the conditional likelihood approach, often used for ARCH-type models. We prove the strong consistency of the conditional likelihood estimators under appropriate conditions. The method is applied to the Kalman filter (for which this contrast and the exact likelihood lead to asymptotically equivalent estimators) and to the discretely observed stochastic volatility models.     

浅析凤翔泥塑对现代设计的启示

中国的民间手工艺文化一直延续着中华本元文化的特征,保持了精神和物质统一的原生状态,同时包蕴了中华民族传统的自然观和思维方式,无论是其中的思想智慧还是造物工艺都为我们的现代设计在如何继承本土文化方面打开了一个可供吸收营养和借鉴的宝库。凤翔泥塑作为一种符号现象被现代设计加以运用,不能不说是民间传统手工艺术在现代的一次回归浪潮。文章探讨的是凤翔泥塑在纹饰、色彩、造型三个方面带给现代设计的启示。     

Mapping water quality by local scoring

The mean density of bacteria in a water body is commonly monitored using quantal assay. This paper describes the use of local scoring in estimating the spatial distribution of mean density from quantal assay results at a set of point locations. An application to estimating the mean density of fecal conform bacteria in a coastal pond is presented. Model diagnostics based on a parametric bootstrap are also presented.