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.     

Blinded continuous monitoring in clinical trials with recurrent event endpoints

In studies with recurrent event endpoints, misspecified assumptions of event rates or dispersion can lead to underpowered trials or overexposure of patients. Specification of overdispersion is often a particular problem as it is usually not reported in clinical trial publications. Changing event rates over the years have been described for some diseases, adding to the uncertainty in planning. To mitigate the risks of inadequate sample sizes, internal pilot study designs have been proposed with a preference for blinded sample size reestimation procedures, as they generally do not affect the type I error rate and maintain trial integrity. Blinded sample size reestimation procedures are available for trials with recurrent events as endpoints. However, the variance in the reestimated sample size can be considerable in particular with early sample size reviews. Motivated by a randomized controlled trial in paediatric multiple sclerosis, a rare neurological condition in children, we apply the concept of blinded continuous monitoring of information, which is known to reduce the variance in the resulting sample size. Assuming negative binomial distributions for the counts of recurrent relapses, we derive information criteria and propose blinded continuous monitoring procedures. The operating characteristics of these are assessed in Monte Carlo trial simulations demonstrating favourable properties with regard to type I error rate, power, and stopping time, ie, sample size.     

The Welfare State as Safety Net in Migration Preferences: Empirical Evidence from an Experiment Among Dutch Master Students

Population Research and Policy Review - The welfare state can be perceived as a safety net which helps individuals adjust to situations of risk or transition. Starting from this idea of the welfare...     

Inequality Assessment by Probabilistic Development Indices

Social Indicators Research - This paper analyses the Human Development Index (HDI) time series from 2010 to 2017. An alternative index is studied, which combines the same components of the HDI by...     

The African Evaluation Guidelines: 2002. A checklist to assist in planning evaluations, negotiating clear contracts, reviewing progress and ensuring adequate completion of an evaluation

A review of the US ‘program evaluation standards’ (PES), undertaken in a series of workshops and meetings of networks of evaluators in Africa, resulted in modifications to those standards. The result was presented to a plenary session of the Inaugural Conference of the African Evaluation Association in September 1999, attended by over 300 evaluators from 35 countries. The AfrEA Conference decided that a systematic effort should be made to produce a list of African evaluation guidelines, similar to the PES, and that this checklist should be reviewed by national evaluation associations and networks in Africa and field tested in several countries. Ten national and regional networks and associations suggested modifications to the text and endorsed the final version of the guidelines.     

Fire and restoration of the largest urban forest of the world in Rio de Janeiro City,Brazil

National Park of Tijuca in Rio de Janeiro (Brazil) is about 3,300 ha and considered the largest urban forest in the world. Its floristic composition is typical of Atlantic Rain Forest. The reserve is being altered because of fire occurrences and urban expansion. This study identified locations and causes of forest fires, and makes management recommendations to restore damaged areas. From 1991 to 2000, forest firefighters recorded an average of 75-fire occurrences/year. Identified causes included hot air balloons (24%), intentional (24%), rubbish burning (21%) and religious practices (17%). Primary fuels included invasive grasses and ferns. Although hot air balloons destroyed larger areas of forest in each occurrence, a greater number of fires started in the invasive vegetation along roads that bisect the forest. In response to recurrent forests, invasive vegetation has spread gradually into the forest increasing forest degradation. To decrease fire damage, sites with high fire frequencies and density of invasive vegetation were planted with less flammable species. Results indicate that fire frequency decreased and density of invasive vegetation declined. This approach appears to prevent fire incidence, reduce the need for fire fighting, and preserve existing biodiversity.     

改善开发所致移民安置项目的后果——牛津大学难民研究中心研究成果报告

论文在讨论四个移民典型案例的基础上 ,讨论在项目中与政策相关的教训     

Form Perception of Partly Occluded Shapes in 4‐Month‐Old Infants

Two habituation experiments were conducted to investigate how 4‐month‐old infants perceive partly occluded shapes. In the first experiment, we presented a simple, partly occluded shape to the infants until habituation was reached. Then we showed either a probable completion (one that would be predicted on the basis of both local and global cues) or an improbable completion. Longer looking times were found for the improbably completed shape (compared to probable and control conditions), suggesting that the probable shape was perceived during partial occlusion. In the second experiment, infants were habituated to more ambiguous partly occluded shapes, where local and global cues would result in different completions. For adults, the percept of these shapes is usually dominated by global influences. However, after habituation the infants looked longer at the globally completed shapes. These results suggest that by the age of 4 months, infants are able to infer the perceptual completion of partly occluded shapes, but for more ambiguous shapes, this completion seems to be dominated by local influences.     

Estimating household structure in ancient China by using historical data: a latent class analysis of partially missing patterns

Summary.  Social data often contain missing information. The problem is inevitably severe when analysing historical data. Conventionally, researchers analyse complete records only. Listwise deletion not only reduces the effective sample size but also may result in biased estimation, depending on the missingness mechanism. We analyse household types by using population registers from ancient China (618–907 AD) by comparing a simple classification, a latent class model of the complete data and a latent class model of the complete and partially missing data assuming four types of ignorable and non-ignorable missingness mechanisms. The findings show that either a frequency classification or a latent class analysis using the complete records only yielded biased estimates and incorrect conclusions in the presence of partially missing data of a non-ignorable mechanism. Although simply assuming ignorable or non-ignorable missing data produced consistently similarly higher estimates of the proportion of complex households, a specification of the relationship between the latent variable and the degree of missingness by a row effect uniform association model helped to capture the missingness mechanism better and improved the model fit.