Robust signed-rank estimation and variable selection for semi-parametric additive partial linear models

A fully nonparametric model may not perform well or when the researcher wants to use a parametric model but the functional form with respect to a subset of the regressors or the density of the errors is not known. This becomes even more challenging when the data contain gross outliers or unusual observations. However, in practice the true covariates are not known in advance, nor is the smoothness of the functional form. A robust model selection approach through which we can choose the relevant covariates components and estimate the smoothing function may represent an appealing tool to the solution. A weighted signed-rank estimation and variable selection under the adaptive lasso for semi-parametric partial additive models is considered in this paper. B-spline is used to estimate the unknown additive nonparametric function. It is shown that despite using B-spline to estimate the unknown additive nonparametric function, the proposed estimator has an oracle property. The robustness of the weighted signed-rank approach for data with heavy-tail, contaminated errors, and data containing high-leverage points are validated via finite sample simulations. A practical application to an economic study is provided using an updated Canadian household gasoline consumption data.     

Quasi-binomial zero-inflated regression model suitable for variables with bounded support

In recent years, a variety of regression models, including zero-inflated and hurdle versions, have been proposed to explain the case of a dependent variable with respect to exogenous covariates. Apart from the classical Poisson, negative binomial and generalised Poisson distributions, many proposals have appeared in the statistical literature, perhaps in response to the new possibilities offered by advanced software that now enables researchers to implement numerous special functions in a relatively simple way. However, we believe that a significant research gap remains, since very little attention has been paid to the quasi-binomial distribution, which was first proposed over fifty years ago. We believe this distribution might constitute a valid alternative to existing regression models, in situations in which the variable has bounded support. Therefore, in this paper we present a zero-inflated regression model based on the quasi-binomial distribution, taking into account the moments and maximum likelihood estimators, and perform a score test to compare the zero-inflated quasi-binomial distribution with the zero-inflated binomial distribution, and the zero-inflated model with the homogeneous model (the model in which covariates are not considered). This analysis is illustrated with two data sets that are well known in the statistical literature and which contain a large number of zeros.     

A COMPARISON OF SEQUENCING RULES FOR A TWO-STATE HYBRID FLOW SHOP

A common manufacturing environment in many industries (such as the glass, steel, paper, costume jewelry, and textile industries) is a hybrid flow shop. This system has continuous-process machinery in the fist state of manufacturing and repetitive-batch equipment in the second. Little research has investigated this type system. Scheduling managers of hybrid flow shops tend either to use existing job-shop rules or to devise their own rules. These approaches often are less than adequate for efficient scheduling. In this paper we extend the rule presented by Narasimhan and Panwalker [4] to include a general class of hybrid flow shops. This extenstion, called the generalized cumulative minimum-deviation (GCMD) rule, is compared under various operation conditions to three other sequencing rules: shortest processing time, longest processing time, and minimum deviation. The operating conditions are determined by the number of machines at both stages. The results of 7200 simulation runs demonstrate that the GCMD rule is better than the other rules in minimizing each of five chosen criteria. Thus, the GCMD rule can help managers to schedule hybrid flow shops efficiently to achieve various corporate objectives.     

VALIDATION OF EXPERT SYSTEMS- WITH APPLICATIONS TO AUDITING AND ACCOUNTING EXPERT SYSTEMS*

This paper proposes a set of definitions for the concepts “validation” and “assessment” applied to expert systems (ESs). It develops a framework for this validation and demonstrates the framework on existing accounting and auditing ESs to elicit some of the research issues involved in ES validation. Validation is critical to the design and implementation of decision-making ESs. In a setting where objectivity is sought and variance is avoided, validation ascertains what a system knows, knows incorrectly, or does not know. Validation ascertains the system's level of expertise and investigates the theoretical basis on which the system is based. It evaluates the reliability of decisions made by the system. The validation framework developed in this paper is research methods. It is designed to reflect the unique aspects of ESs (in contrast to other types of computer programs) and can be used by ES developers as a basis from which to perform validation and by researchers as a framework to elicit research issues in validation.     

Notes and Communications “A DECISION THEORY EXAMPLE IN FOOTBALL”: A COMMENT*

Janssen and Daniel analyzed the choice between a one- or a two-point conversion for a particular game situation in college football. Their decision criteria was maximum expected utility based on a von Neumann-Morgenstern utility function defined over the games outcomes. An alternative approach based on a stochastic dominance criterion is presented that does not rely on knowledge of the relative importance of tying vs. winning; rather, it relies on a notion of consistency in the sequential problem.     

A NOTE ON HOLDING COSTS AND LOT-SIZE ERRORS

This note examines the sensitivity of the basic economic-order-quantity inventory model to lot-size errors when holding costs are assumed to be a strictly increasing (though not necessarily linear) function of average inventory. In particular, we show that the penalty associated with ordering either too much or too little is a function not only of the size of the error but of the shape of the holding-cost curve as well. We demonstrate that, under certain conditions, even relatively small lot-size errors can be extremely costly.