Showing papers on "Model selection published in 1980"
TL;DR: In this paper, the basic ideas underlying the construction of a newly introduced seasonal adjustment procedure by a Bayesian modeling are discussed in detail, with particular emphasis on the use of the concept of the likelihood of a bayesian model for model selection.
Abstract: The basic ideas underlying the construction of a newly introduced seasonal adjustment procedure by a Bayesian modeling are discussed in detail. Particular emphasis is placed on the use of the concept of the likelihood of a Bayesian model for model selection. The performance of the procedure is illustrated by a numerical example.
149 citations
TL;DR: In this article, the authors evaluate the performance of a representative group of smoothing models over a variety of conditions in 9,000 simulated time series and demonstrate that a major disadvantage of adaptive smoothing model is their tendency to generate unstable forecasts, even during periods when mean demand itself is stable.
Abstract: Despite the general acceptance of exponential smoothing, the choice of a specific smoothing model is often a difficult problem. Previous research involving smoothing-model comparisons and the penalties for selection of the wrong model has been limited. This paper evaluates the performance of a representative group of smoothing models over a variety of conditions in 9,000 simulated time series. Forecast-error results demonstrate that a major disadvantage of adaptive smoothing models is their tendency to generate unstable forecasts, even during periods when mean demand itself is stable. Several trend-adjusted smoothing models are shown to be robust forecasters, whether the time series actually display a trend or not.
97 citations
TL;DR: A recent paper by Benedetti and Brown as mentioned in this paper suggests that a formal simultaneous test may be applied to the entire model selection procedure, including the testing of models chosen after inspection of the data.
Abstract: A recent paper in the Consultant's Forum by Benedetti and Brown (1978) discusses possible strategies for the choice of an appropriate log-linear model for a comples contingency table. The methods considered do not invariably choose the same model, but the authors favour the tests of marginal and partial association due to Brown (1976). The authors point out difficulties with the simultaneous test of all pooled effects of the same order and, in a final sentence, note that the significance levels [of multiple tests of the data] are merely guidelines and should not be interpreted as formal tests of significance.' It is the purpose of this note to point out that a formal simultaneous test may be applied to the entire model selection procedure, including the testing of models chosen after inspection of the data. The simultaneous test will generally use Type I error rates that are much higher than those typically used in discussions of model choice. The choice of a parsimonious model may then become a systematic procedure generally requiring only a small number of explicit fits of models.
43 citations
TL;DR: The development of parametric sighting models and estimation techniques is reviewed, and general procedures for variance estimation are given, and two new models and their corresponding abundance estimators are developed.
Abstract: The estimation of the abundance of animal populations with the line—transect sampling method requires that a sighting model be postulated to represent the decreased probability of a sighting as its distance from the transect line increases. The development of parametric sighting models and estimation techniques is reviewed, and general procedures for variance estimation are given. Two new models and their corresponding abundance estimators are developed. The half—normal model provides a useful alternative to the widely used exponential model. The generalized exponential model contains the half—normal and exponential models as special cases and has the flexibility to account for a wide range of sighting phenomena. The new models are applied to transect surveys of porpoise schools and the results are related to sampling and ecological conditions. Criteria for model selection and transect design are discussed. See full-text article at JSTOR
33 citations
TL;DR: There is no single, best estimate model for predicting critical two-phase flow mass rates, and though comparisons of model predictions with experimental data provide an interim guide, much more work is needed to resolve acceptability of models.
17 citations
TL;DR: In this article, a selection process on a finite number of individuals divided into K classes is considered, and the individuals are selected one by one, without l-eplacemellt, until a fixed number has been chosen.
Abstract: A selection process on a finite number of individuals divided into K classes is considered. The individuals are selected one by one, without l-eplacemellt, until a fixed number has been chosen. It is assumed that individuals in the same class have the same probability of being chosen at any stage, but this probability can vary from class to class. More precisely, the probability of selection from a class is assumed to be proportional to the product of the number of individuals remaining in that class and a weighting factor. This model appeals to be quite realistic for some types of selective predation experiments. Equations are provided for the maximum likelihood estimation of the weighting factors, on the assumption that the order of choice of individuals is known. Simple approximate expressions for the variances and covariances are also given. A simulation experiment has shown that the estimation of the weighting factors is virtually unbiased. The approximate expressions for variances and covariances also appear to be quite accurate for a wide variety of situations. For tests of significance and confidence limits a logarithmic transformation has been found to be worthwhile.
9 citations
TL;DR: Nonlinear techniques for model selection are used and compared with traditional techniques and interactive graphics techniques are used to advantage for evaluating various models and for examining the quality of the experimental data.
Abstract: The use of an automated reactor for the experimentation and on-line graphics for the rapid and exhaustive analysis of experimental data is described. Traditional (linear) methods are used for selecting the most promising model for the alkaline phosphatase catalyzed reaction from a set of ten models under consideration. Then, nonlinear techniques for model selection are used and compared with traditional techniques. In both approaches, interactive graphics techniques are used to advantage for evaluating various models and for examining the quality of the experimental data.
8 citations
TL;DR: In this article, three model selection strategies were developed, namely, a strategy to be used in anticipation of large coefficients of variation (neighborhood of 65 percent), a strategy for small coefficients of variations (4 percent or less), and a security regret strategy for the absence of prior knowledge.
Abstract: As many as three iterated statistical model deletion procedures are considered for an experiment.Population model coeff cients were chosen to simulate a saturated 24experiment having an unfavorable distribution of parameter values.Using random number studies, three model selection strategies were developed, namely, (1) a strategy to be used in anticipation of large coefficients of variation (neighborhood of 65 percent), (2) strategy to be used in anticipation of small coefficients of variation (4 percent or less), and (3) a security regret strategy to be used in the absence of such prior knowledge
1 citations
1 citations
01 Jan 1980
TL;DR: In this paper, identification and model selection in econometrics are examined from a system theoretic point of view, and a critical study of the existing techniques in Econometric is carried out through several examples from literature.
Abstract: : Identification and model selection in econometrics are examined from a system theoretic point of view. A critical study of the existing techniques in econometrics is carried out through several examples from literature. (Author)
TL;DR: This paper examined the test performance of 100 8- to 11-year-old children on subtraction tasks involving variations in borrowing and found that tasks including various types of borrowing were equivalent and that the ability to substract without borrowing was prerequisite to subtraction involving borrowing.