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.     

Large scale communication networks—Design and analysis

The paper discusses the problem of design and analysis of large scale communication systems. An iterative process, composed of a minimum cost network design algorithm and a network performance algorithm, is presented for the solution of these problems. Computational considerations, using these algorithms, are discussed.     

An exploratory study of the influence of the IS function and organizational context on business process reengineering project initiatives

To help evaluate the risk of process reengineering failure and enhance the prospect of its success, three potential sources of influence on BPR initiatives and success are examined in this study. These include the innovative capacity of the organization, IS maturity and strategy-IS interface. It was found that while factors related to IT maturity and influence such as experience in mainframe and client/server computing may facilitate the decision to reengineer, they are not critical in the later stages of the initiative. On the other hand, factors having significant relationships beyond the initial decision include variables pertaining to innovative capacity of the organization and strategy-IS interface. These findings suggest that technical IT competence as a critical enabler is necessary but not sufficient for reengineering success. Based on study findings regarding the innovative capacity of the organization, guidelines for reengineering risk assessment are proposed. In addition, implications of the findings, limitations of the study and opportunities for further research are also discussed.     

Influencing alcohol control policies and practices at community festivals

The goal of this study was to assess the feasibility and effectiveness of two interventions aimed at reducing alcohol-related risks at community festivals--a training program for festival planners and a community organizing campaign. We randomly selected four festivals for each intervention and had 24 comparison festivals. Our assessment included process evaluation to track and evaluate types of alcohol policies resulting from each of the interventions, pre and post telephone surveys of key festival planners and law enforcement agencies, and pre and post pseudo-underage and pseudo-intoxicated purchase attempts. Analyses showed that both interventions were feasible and were successful in influencing adoption of written policies and improving alcohol-related practices. However, neither intervention appeared to decrease propensity for illegal alcohol sales at these events, likely due, in part, to the short time frame of the interventions. Future research should assess effects of the interventions on alcohol-related problems and effects of enforcement interventions.     

Handheld computers. A feasible alternative to paper forms for field data collection

Recent advances in handheld computer hardware and software may provide alternatives to paper-based data collection methods. The authors compared data collected with paper forms to data collected with handheld computer-based forms in a field observation study of alcohol purchase attempts at 47 community festivals in a large metropolitan area. Agreement between data collected with paper forms and data collected with handheld computers was greater than 95%. Computer-based forms handled branching patterns better and yielded data that were immediately available for analyses. Paper forms handled written comments better. Handheld computers are a feasible alternative to paper forms for field data collection.     

Effectiveness and Efficiency of Controlling Campylobacter on Broiler Chicken Meat

Campylobacter bacteria are an important cause of foodborne infections. We estimated the potential costs and benefits of a large number of possible interventions to decrease human exposure to Campylobacter by consumption of chicken meat, which accounts for 20-40% of all cases of human campylobacteriosis in the Netherlands. For this purpose, a farm-to-fork risk assessment model was combined with economic analysis and epidemiological data. Reduction of contamination at broiler farms could be efficient in theory. However, it is unclear which hygienic measures need to be taken and the costs can be very high. The experimental treatment of colonized broiler flocks with bacteriophages has proven to be effective and could also be cost efficient, if confirmed in practice. Since a major decrease of infections at the broiler farm is not expected in the short term, additional measures in the processing plant were also considered. At this moment, guaranteed Campylobacter-free chicken meat at the retail level is not realistic. The most promising interventions in the processing plant are limiting fecal leakage during processing and separation of contaminated and noncontaminated flocks (scheduling), followed by decontamination of the contaminated flock. New (faster and more sensitive) test methods to detect Campylobacter colonization in broilers flocks are a prerequisite for successful scheduling scenarios. Other methods to decrease the contamination of meat of colonized flocks such as freezing and heat treatment are more expensive and/or less effective than chemical decontamination.     

Modular Network Design: Using Information and Communication Technology to Allocate Production Tasks in a Virtual Organization*

The increased interest in customer service and the trend towards customization gives rise to new organizational forms such as the virtual organization. This paper introduces the Modular Network Design (MND) approach to complement Mowshowitz' (1997) theory of virtual organization. This theory focuses on metamanagement of the virtual organization and consists of four activities: (1) determine and analyze customer requirements, (2) track the possibilities for satisfying these requirements, (3) develop and allocate production tasks among the members of the virtual organization, and (4) assess and adjust tasks and allocation procedures. The MND approach elaborates on this theory by breaking down both the requirements and the production tasks into modular entities, and by measuring the performance of the resulting virtual network in terms of operating costs and throughput time. The major contribution of the MND approach lies in its ability to support the assessment of alternative allocations of production tasks among the members of a virtual organization, using Information and Communication Technology (ICT) as enabler. The assessment of such alternative designs is illustrated by an application of MND in the air cargo industry. The results show how the virtual organization and its members can realize significant improvements in operating costs and throughput times when applying ICT to link and reallocate their modular production tasks in response to customized orders.     

Assessing Interventions to Reduce the Risk of Campylobacter Prevalence in Broilers

As part of a comprehensive risk assessment on the Campylobacter prevalence in the chicken production chain (from young born chicken till chicken fillet) in the Netherlands, we formulated a quantitative model on the transmission dynamics of Campylobacter at Dutch broiler farms. This model is used to quantify the risk of Campylobacter prevalence in broilers at the time that flocks leave the farm for processing. To this end, we assumed that the Campylobacter prevalence is primarily determined by two parameters, that is, the within- and between-flock transmission. The within-flock transmission was assessed fitting experimental data to a logistic growth model and the between-flock transmission was assessed fitting field data to a generalized linear model (GLM), which included three possible infection routes: (1) via an infected flock in the previous cycle, (2) via other infected flocks present on the farm, and (3) from other sources. This model was applied to assess the efficacy of three control scenarios; (1) a ban on other livestock on broiler farms, (2) a ban on thinning, and (3) a reduction of the between-flock transmission. In contrast to the other scenarios, the third one was shown to be most effective. Theoretically, this is accomplished by improved biosecurity. However, the impact of improved biosecurity cannot be specified into specific control measures, and therefore it is not clear what investments are needed. Finally, we also assessed the efficacy of scheduled treatment, that is, fresh meat production solely from test-negative flocks. We found that the reliability of negative test results, which is crucial, strongly depends on the length of time between testing and slaughter. The sensitivity and specificity of the test appeared to be of minor importance.     

Accounting, strategy and AMT investment

One strand in the current debate on AMT investment emphasises technical aspects of investment appraisal. Another calls for a broader focus on wider decision-making processes and strategic considerations. This paper seeks to merge both sets of concerns. It uses two empirical studies of AMT decisions in a number of companies. The first demonstrates specific linkages between manufacturing strategies and financial justifications. The second explores ways in which connections between finance and strategy are constructed and how they are understood and evaluated by decision-makers. We conclude that AMT proposals succeed by linking accounting information to strategic considerations throughout the investment decision process.