Model Risk

A cross-country pipeline construction project is exposed to an uncertain environment due to its enormous size (physical, manpower requirement and financial value), complexity in design technology and involvement of external factors. These uncertainties can lead to several changes in project scope during the process of project execution. Unless the changes are properly controlled, the time, cost and quality goals of the project may never be achieved. A methodology is proposed for project control through risk analysis, contingency allocation and hierarchical planning models. Risk analysis is carried out through the analytic hierarchy process (AHP) due to the subjective nature of risks in construction projects. The results of risk analysis are used to determine the logical contingency for project control with the application of probability theory. Ultimate project control is carried out by hierarchical planning model which enables decision makers to take vital decisions during the changing environment of the construction period. Goal programming (GP), a multiple criteria decision-making technique, is proposed for model formulation because of its flexibility and priority-base structure. The project is planned hierarchically in three levels—project, work package and activity. GP is applied separately at each level. Decision variables of each model are different planning parameters of the project. In this study, models are formulated from the owner's perspective and its effectiveness in project control is demonstrated.

Abstract: We consider the problem of variable selection for logistic regression when the dependent variable is measured imperfectly, under both differential and non-differential misclassification. An MCMC sampling scheme is designed, incorporating uncertainty ...

The International Life SciencesInstitute (ILSI) Risk Science Institute (RSI) convenedan expert panel of scientists to developrecommendations for a comprehensive monitoring programfor the Croton and Catskill/Delaware watersheds, whichprovide drinking water to New York City's residents. This effort was conducted as part of efforts topreserve and enhance the quality of New York City'sreservoir system through a watershed protectionprogram. The panel developed recommendations

The recent financial crisis has raised numerous questions about the accuracy of value-at-risk (VaR) as a tool to quantify extreme losses. In this paper we present empirical evidence from assessing the out-of-sample performance and robustness of VaR before and during the recent financial crisis with respect to the choice of sampling window, return distributional assumptions and stochastic properties of the

A theory-based protection/risk model was applied to explain variation in college students' heavy episodic drinking. Key aims were (1) to establish that psychosocial and behavioral protective factors and risk factors can account for cross-sectional and developmental variation in heavy episodic drinking, and (2) to examine whether protection moderates the impact of risk on heavy episodic drinking. Random- and fixed-effects maximum likelihood regression analyses were used to examine data from a three-wave longitudinal study. Data were collected in fall of 2002, spring of 2003, and spring of 2004 from college students (N=975; 548 men) who were first-semester freshmen at Wave 1. Psychosocial and behavioral protective and risk factors accounted for substantial variation in college-student heavy episodic drinking, and protection moderated the impact of risk. Findings held for both genders and were consistent across the three separate waves of data. Key predictors of heavy episodic drin...