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.     

The value of being precise

This study investigates how individuals assess imprecise information. We focus on two essential dimensions of decision under uncertainty, outcomes and probabilities, and their respective precision. We believe the precision of information is highly relevant in the investment setting, as reflected in the well-known “home (familiarity) bias”, and the outcome and probability dimensions, separately or jointly, may affect investors’ knowledge of uncertainty and perceived risk of the investment options, and subsequently affect investors’ choices. To test this conjecture, we conducted three experiments. Our results show that 1) participants demonstrate a pattern of preference for precision and aversion of extreme vagueness and associate vagueness with higher perceived risk and lower investment (experiments one and two); 2) participants prefer vague outcome information to vague probability information (experiment two); 3) familiarity indeed positively affects the precision of estimated values, but this association is stronger for the outcome dimension than for probabilities (experiment three). Our results confirm that precision in information, especially in the outcome dimension has an impact on investors’ resource allocation choices.     

Price bias and common practice in option pricing

Generally, the semiclosed-form option pricing formula for complex financial models depends on unobservable factors such as stochastic volatility and jump intensity. A popular practice is to use an estimate of these latent factors to compute the option price. However, in many situations this plug-and-play approximation does not yield the appropriate price. This article examines this bias and quantifies its impacts. We decompose the bias into terms that are related to the bias on the unobservable factors and to the precision of their point estimators. The approximated price is found to be highly biased when only the history of the stock price is used to recover the latent states. This bias is corrected when option prices are added to the sample used to recover the states' best estimate. We also show numerically that such a bias is propagated on calibrated parameters, leading to erroneous values. The Canadian Journal of Statistics 48: 8–35; 2020 © 2019 Statistical Society of Canada     

Visualization of generalized mean estimators using auxiliary information in survey sampling

Abstract

The mean estimators with ratio depend on multiple auxiliary variables and unknown parameters in a finite population setting. We propose a new generalized approach with matrices for modeling the mutivariate mean estimators with two auxiliary variables. Our approach brings naturally a graphical analysis for comparing mean estimators.     

Product partition latent variable model for multiple change-point detection in multivariate data

The product partition model (PPM) is a well-established efficient statistical method for detecting multiple change points in time-evolving univariate data. In this article, we refine the PPM for the purpose of detecting multiple change points in correlated multivariate time-evolving data. Our model detects distributional changes in both the mean and covariance structures of multivariate Gaussian data by exploiting a smaller dimensional representation of correlated multiple time series. The utility of the proposed method is demonstrated through experiments on simulated and real datasets.     

政府减排目标、产业结构与污染减排

政府是污染减排的重要主体,了解其行为对污染减排的作用对提升污染减排效果及实现中国绿色发展具有重要意义。创造性地将政府减排目标纳入研究框架,构建面板门槛模型,从工业污染排放总量的角度分析政府减排目标、产业结构、经济规模等因素对地方污染减排的影响,并用工业污染物强度进行稳健性检验。结果显示,政府减排目标对污染减排的作用存在阶段性不同:经济发展初级阶段,政府减排目标对污染排放总量的作用方向为正; 经济发展水平较高时,作用方向为负; 两者之间存在适应性调整阶段,作用方向具有不确定性,但总体由正向作用向负向作用转变。产业结构对工业污染排放总量的作用方向因污染物种类而异,经济规模对工业污染排放总量作用方向为正。     

市场分割、政企关系与异地市场拓展

中国的市场仍旧存在某种程度的分割，如何突破这种市场分割到异地拓展市场是企业要考虑的现实问题。以前的研究主要是从省级层面或国家层面进行宏观分析。文章创新地从政企关系视角入手来考察市场分割情形下微观层面的企业异地市场拓展，也是首先把市场分割（和异地市场）分为省内（本市外）市场、省外市场和海外市场三个不同层面来研究的论文。用世界银行的调查数据从微观企业个体层面检验了企业异地市场拓展的可能。实证研究结果发现，市场分割确实阻碍了企业异地市场拓展，但这仅限于省内市场而不是省外市场。此外，市场分割是有方向性的，从地方保护主义严重的市场里走出来的企业更容易拓展异地市场，但此结论仅适用于国内市场拓展而不适用于海外市场拓展。良好的政企关系是有利于企业异地市场拓展的，但仅限于省外市场，对省内市场以及海外市场是无效的。文章解释了其中的原因，并分析了对于企业实践的启示。     

虚拟空间交织的取向原则

虚拟空间交织是人们常用的思维与认知方式之一。本文通过对交织及其原则的介绍,描述了虚拟空间交织的运作,指出虚拟空间交织构建中主观取向的原则应为最佳关联。     

ALL-BIAS DESIGNS FOR POLYNOMIAL SPLINE REGRESSION MODELS

Polynomial spline regression models of low degree have proved useful in modeling responses from designed experiments in science and engineering when simple polynomial models are inadequate. Where there is uncertainty in the number and location of the knots, or breakpoints, of the spline, then designs that minimize the systematic errors resulting from model misspecification may be appropriate. This paper gives a method for constructing such all‐bias designs for a single variable spline when the distinct knots in the assumed and true models come from some specified set. A class of designs is defined in terms of the inter‐knot intervals and sufficient conditions are obtained for a design within this class to be all‐bias under linear, quadratic and cubic spline models. An example of the construction of all‐bias designs is given.