In this study, we propose a control chart for monitoring two-stage processes whose quality characteristic to be monitored in the second stage follows a binomial distribution. The proposed control chart is based on the...
详细信息
In this study, we propose a control chart for monitoring two-stage processes whose quality characteristic to be monitored in the second stage follows a binomial distribution. The proposed control chart is based on the deviance residual in which essentially the generalized log-likelihood ratio statistic is obtained from the generalized linear model. To establish the relationship between the first- and second-stage quality characteristics, we propose using a new link function in a generalized linear model framework. The performance of the proposed control chart with the new link function is compared with that under the traditional logit link function in terms of the average run length criterion. In addition, the performance of the proposed control chart is compared with the chart designed based on the original residuals under the new link function as well as the traditional np-chart applied for monitoring the binomial quality characteristic in the second stage.
The intention of this study was to identify a suitable generalized linear model (GLM) for modelling multi-site daily rainfall in the Onkaparinga catchment in South Australia and to examine the suitability of the model...
详细信息
The intention of this study was to identify a suitable generalized linear model (GLM) for modelling multi-site daily rainfall in the Onkaparinga catchment in South Australia and to examine the suitability of the model for downscaling of General Circulation model (GCM) rainfall projections. A GLM was applied and multi-site daily rainfall was downscaled using National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis datasets. Nineteen large-scale atmospheric and circulation variables were selected at first and these were eventually reduced, based on correlation with daily rainfall, to 10 final variables to be used in the model. First, logistic regression was used to identify the wet and dry days, then wet day rainfall was modelled using a gamma distribution. The model was fitted for a calibration period (1991-2010) and it was then validated over the period 1981-1990. Several summary statistics including mean, standard deviation, number of wet days, maximum rainfall amount and lag 1 and lag 2 autocorrelations were used to check the model performance. The 2.5th and 97.5th percentiles of the simulated rainfall statistics were plotted against the observed rainfall statistics and it was shown that most of the observed statistics were within these bounds. Area averaged and station wise monthly, seasonal and annual totals for observed and simulated rainfall were estimated and compared. The overall performance of the GLM to downscale rainfall was considered satisfactory. However, a few discrepancies were observed in different performance statistics. Parameterization of the model to capture the local convective variability of rainfall would increase the model performance. It was found overall that the GLM can be applied for downscaling of GCM rainfall projections for this catchment.
作者:
Feng, JianfengGao, YongfeiJi, YijunZhu, LinNankai Univ
Key Lab Pollut Proc & Environm Criteria Minist Educ Coll Environm Sci & Engn Tianjin 300071 Peoples R China Nankai Univ
Tianjin Key Lab Environm Remediat & Pollut Contro Coll Environm Sci & Engn Tianjin 300071 Peoples R China Chinese Acad Agr Sci
Inst Grassland Res Hohhot 010010 Peoples R China
Predicting the toxicity of chemical mixtures is difficult because of the additive, antagonistic, or synergistic interactions among the mixture components. Antagonistic and synergistic interactions are dominant in meta...
详细信息
Predicting the toxicity of chemical mixtures is difficult because of the additive, antagonistic, or synergistic interactions among the mixture components. Antagonistic and synergistic interactions are dominant in metal mixtures, and their distributions may correlate with exposure concentrations. However, whether the interaction types of metal mixtures change at different time points during toxicodynamic (TD) processes is undetermined because of insufficient appropriate models and metal bioaccumulation data at different time points. In the present study, the generalized linear model (GLM) was used to illustrate the combined toxicities of binary metal mixtures, such as Cu-Zn, Cu-Cd, and Cd-Pb, to zebrafish larvae (Danio rerio). GLM was also used to identify possible interaction types among these method for the traditional concentration addition (CA) and independent action (IA) models. Then the GLM were applied to quantify the different possible interaction types for metal mixture toxicity (Cu-Zn, Cu-Cd, and Cd-Pb to D. rerio and Ni-Co to Oligochaeta Enchytraeus crypticus) during the TD process at different exposure times. We found different metal interaction responses in the TD process and interactive coefficients significantly changed at different exposure times (p<0.05), which indicated that the interaction types among Cu-Zn, Cu-Cd, Cd-Pb and Ni-Co were time dependent. Our analysis highlighted the importance of considering joint actions in the TD process to understand and predict metal mixture toxicology on organisms. Moreover, care should be taken when evaluating interactions in toxicity prediction because results may vary at different time points. The GLM could be an alternative or complementary approach for BLM to analyze and predict metal mixture toxicity. (C) 2017 Elsevier B.V. All rights reserved.
The penalized variable selection methods are often used to select the relevant covariates and estimate the unknown regression coefficients simultaneously,but these existing methods may fail to be consistent for the se...
详细信息
The penalized variable selection methods are often used to select the relevant covariates and estimate the unknown regression coefficients simultaneously,but these existing methods may fail to be consistent for the setting with highly correlated *** this paper,the semi-standard partial covariance(SPAC)method with Lasso penalty is proposed to study the generalized linear model with highly correlated covariates,and the consistencies of the estimation and variable selection are shown in high-dimensional settings under some regularity *** simulation studies and an analysis of colon tumor dataset are carried out to show that the proposed method performs better in addressing highly correlated problem than the traditional penalized variable selection methods.
Multicollinearity problem arises frequently in several modern applications, such as chemometrics, biology, and other scientific fields. The common feature of the multicollinearity problem is that a large number of pre...
详细信息
Multicollinearity problem arises frequently in several modern applications, such as chemometrics, biology, and other scientific fields. The common feature of the multicollinearity problem is that a large number of predictors are highly correlated. generalized linear model is a powerful and a popular approach for modeling a large variety of regression data. It is well known that the existence of multicollinearity can inflate the variance of the maximum likelihood estimator. To reduce the effects of multicollinearity, the ridge estimator has been efficiently demonstrated to be an attractive method. However, the choice of the biasing parameter of the ridge estimator is critical. Our aim is to efficiently estimate such a biasing parameter. Towards this aim, a kidney-inspired algorithm, which is a population-based algorithm inspiring by the kidney process in the human body, is proposed. Extensive comparisons with different classical biasing parameter estimating methods are conducted through simulation and real data application. The results demonstrate that our proposed approach is able to find the best biasing parameter value with high prediction accuracy. Further, the results indicate that the performance of our proposed approach is superior to that of other competitor methods.
作者:
MORTON, RCSIRO
DIV MATH & STAT CANBERRA ACT 2601 AUSTRALIA
We consider a scaled Poisson generalized linear model with random multiplicative errors associated with each stratum of a nested block structure. The normal equations are obtained. The quasi-likelihood is shown to be ...
详细信息
We consider a scaled Poisson generalized linear model with random multiplicative errors associated with each stratum of a nested block structure. The normal equations are obtained. The quasi-likelihood is shown to be the product of a sequence of negative binomial quasi-likelihoods using weighted totals and correcting for margins. This is achieved by extending the definition of the quasi-likelihood which now includes a vector of weights. The usual asymptotic results still hold. Data on trap catches of insects provide a numerical example.
Recently, the selection consistency of penalized least square estimators has received a great deal of attention. For the penalized likelihood estimation with certain non-convex penalties, search space can be construct...
详细信息
Recently, the selection consistency of penalized least square estimators has received a great deal of attention. For the penalized likelihood estimation with certain non-convex penalties, search space can be constructed within which there exists a unique local minimizer that exhibits selection consistency in high-dimensional generalized linear models under certain conditions. In particular, we prove that the SCAD penalty of Fan and Li (2001) and a new modified version of the unbounded penalty of Lee and Oh (2014) can be employed to achieve such a property. These results hold even for the non-sparse cases where the number of relevant covariates increases with the sample size. Simulation studies are provided to compare the performance of SCAD penalty and the newly proposed penalty. (C) 2016 Elsevier B.V. All rights reserved.
There has been a considerable amount of work devoted by transportation safety analysts to the development and application of new and innovative models for analyzing crash data. One important characteristic about crash...
详细信息
There has been a considerable amount of work devoted by transportation safety analysts to the development and application of new and innovative models for analyzing crash data. One important characteristic about crash data that has been documented in the literature is related to datasets that contained a large amount of zeros and a long or heavy tail (which creates highly dispersed data). For such datasets, the number of sites where no crash is observed is so large that traditional distributions and regression models, such as the Poisson and Poisson-gamma or negative binomial (NB) models cannot be used efficiently. To overcome this problem, the NB-Lindley (NB-L) distribution has recently been introduced for analyzing count data that are characterized by excess zeros. The objective of this paper is to document the application of a NB generalized linear model with Lindley mixed effects (NB-L GLM) for analyzing traffic crash data. The study objective was accomplished using simulated and observed datasets. The simulated dataset was used to show the general performance of the model. The model was then applied to two datasets based on observed data. One of the dataset was characterized by a large amount of zeros. The NB-L GLM was compared with the NB and zero-inflated models. Overall, the research study shows that the NB-L GLM not only offers superior performance over the NB and zero-inflated models when datasets are characterized by a large number of zeros and a long tail, but also when the crash dataset is highly dispersed. Published by Elsevier Ltd.
Remote analysis systems allow analysts to obtain statistical results without providing direct access to confidential data stored in a secure server system. An attacking analyst could send queries to a remote server to...
详细信息
Remote analysis systems allow analysts to obtain statistical results without providing direct access to confidential data stored in a secure server system. An attacking analyst could send queries to a remote server to obtain outputs of statistical analyses and use those outputs for a disclosure attack. Statistical disclosure control (SDC) methods are used to modify remote analysis system (RAS) outputs in the protection of confidential information. Confidentiality protection through perturbation is one of the most commonly adopted SDC methods. In the case of generalized linear modelling, random noise is added to the estimated coefficients or to the associated estimating equation prior to getting estimates. This inflates the variances of estimators, and some efficiency and utility of estimators are lost. Thus the application of any perturbation based SDC method could result in an inefficient estimator, with the danger of producing worthless inferences. To date, little attention has been given to systematically controlling the disclosure risk and utility in SDC methods for RAS. In this paper, we develop a framework for the perturbation of estimating equations that enables an RAS to release modified generalized linear model output in such a way that the disclosure risk is not only reduced but also a good utility is maintained. Finally, we present some empirical results demonstrating the application of our framework for obtaining estimates from perturbed estimating equations of binary and count response models.
We propose to use a penalized estimator for detecting homogeneity of the high dimensional generalized linear model. Here, the homogeneity is a specific model structure where regression coefficients are grouped having ...
详细信息
We propose to use a penalized estimator for detecting homogeneity of the high dimensional generalized linear model. Here, the homogeneity is a specific model structure where regression coefficients are grouped having exactly the same value in each group. The proposed estimator achieves weak oracle property under mild regularity conditions and is invariant to the choice of reference levels when there are categorical covariates in the model. An efficient algorithm is also provided. Various numerical studies confirm that the proposed penalized estimator gives better performance than other conventional variable selection estimators when the model has homogeneity. (C) 2017 Elsevier B.V. All rights reserved.
暂无评论