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.     

Radical Distrust: Are Economic Policy Attitudes Tempered by Social Trust?

Social Indicators Research - Debates about the appropriate role of markets and governments are often shaped by sharply contrasting opinions. Based on individual data from the World Values Survey...     

A vector of quarters representation for bivariate time series

In this paper it is shown that several models for a bivariate nonstationary quarterly time series are nested in a vector autoregression with cointegration restrictions for the eight annual series of quarterly observations. Or, the Granger Representation Theorem is extended to incorporate, e.g., seasonal and periodic cointegration.     

Book reviews

Statistical Papers -     

Detecting seasonal unit roots in a structural time series model

In this paper, we propose to detect seasonal unit roots within the context of a structural time series model. Such a model is often found to be useful in practice. Using Monte Carlo simulations, we show that our method works well. We illustrate our approach for several quarterly macroeconomic time series variables.     

Dramatically reducing cycle times through flash development

The reduction of cycle time has emerged as a fundamental element of new product development strategy. Many companies look solely at speed enhancement and increase the human, capital and technical resources allocated to development projects. However, additional resources do not guarantee significant reductions in cycle time. As well as speed enhancement activities, there exist two other sets of activities that are key to shortening the development cycle: those that act to prevent delays (friction) in the development process and those that define the scope of the final outcome (direction). This paper describes flash development as an alternative approach to reducing product development cycle times. Flash development works by creating a process that focuses on speed, friction and direction. The process is illustrated through an analysis of the Galileo Project at Nortel Networks.