In this paper,we study optimal model averaging estimators of regression coefficients in a multinomial logit model,which is commonly used in many scientific fields.A Kullback-Leibler(KL)loss-based weight choice criteri...In this paper,we study optimal model averaging estimators of regression coefficients in a multinomial logit model,which is commonly used in many scientific fields.A Kullback-Leibler(KL)loss-based weight choice criterion is developed to determine averaging weights.Under some regularity conditions,we prove that the resulting model averaging estimators are asymptotically optimal.When the true model is one of the candidate models,the averaged estimators are consistent.Simulation studies suggest the superiority of the proposed method over commonly used model selection criterions,model averaging methods,as well as some other related methods in terms of the KL loss and mean squared forecast error.Finally,the website phishing data is used to illustrate the proposed method.展开更多
The multiple cause of death (MCOD) analysis is used to account for the full contribution of TB as a cause of death to South African mortality in 2008 that were coded using ICD10. Following a review of MCOD methods, a ...The multiple cause of death (MCOD) analysis is used to account for the full contribution of TB as a cause of death to South African mortality in 2008 that were coded using ICD10. Following a review of MCOD methods, a sufficient set of variables for use in MCOD and a new method of quantifying the severity of each cause of death are proposed. The results show that a total of 86,818 (14.3% of all deaths) were TB related, and within all deaths due to natural underlying causes, 86,373 (16.1%) were TB related. Furthermore, 42,581 (7.9%) were due to TB only, 6.0% had TB as an underlying cause along with other contributory causes and 2.0% had TB as a contributory cause. TB was mentioned as the underlying cause of death in 74,863 certificates or 13.9% of deaths due to natural underlying causes. Further analysis using multinomial baseline logit models, reveals that the relative odds of death in any demographic group compared with death in the baseline categories depend on the severity level of TB considered. It is proposed that the severity measure should be adopted when studying the contribution of all main causes of death to total mortality.展开更多
This study applied multilevel modeling to investigate the impact of observed predictors and different levels or groups that households belong, on parents’ choice of discipline methods using data from 8156 households ...This study applied multilevel modeling to investigate the impact of observed predictors and different levels or groups that households belong, on parents’ choice of discipline methods using data from 8156 households derived from a nationwide survey by the Ghana Statistical Service (GSS) in 2011. The aim of the study is to provide in-depth information on why parents choose particular discipline methods as corrective measures to reduce unwanted child behaviour in the present and to increase desirable ones in the future. The results of the study show that, religion and age-group of household heads have significant effect on household’s likelihood to choose physical discipline methods whereas the wealth index of a household and ethnicity of the household head, have significant effect on households’ likelihood to choose non-physical and psychological aggression methods. The results further show significant contextual effect on the differences in choices of parents at the household and regional levels. The choice of physical discipline methods by parents was consistent across households and regional levels unlike non-physical and psychological aggression methods whose application varied across the regions. Households in the Northern, Eastern and Volta regions mostly chose to apply physical discipline methods whereas in the Upper West, Western and Northern regions the most chosen discipline methods were non-physical discipline methods. Psychological aggression discipline methods were predominantly applied in the Upper East, Central and Northern regions of the country.展开更多
A growing stream of study stresses the relevance of subjective elements in understanding the hierarchy of preferences that underpin individual travel behavior. The purpose of this study is to evaluate the impact of va...A growing stream of study stresses the relevance of subjective elements in understanding the hierarchy of preferences that underpin individual travel behavior. The purpose of this study is to evaluate the impact of various factors on mode choice. To achieve this, a multinomial logit model (MNL) was used to analyze the relationships between mode choice and three classes of attributes;Combined Active and Latent, Active only and Latent only attributes. The data used are derived from surveys in the port city of Douala, Cameroon as a case study. Results stipulated that, the combined attributes model performed better than both active only attributes and latent only attributes models. Likewise, latent only attributes model performed better than active only attributes model. The advantage of modelling all three groups is for better selection of the most relevant attributes, and this is very relevant in understanding travel behavior of individuals and mode choice decisions.展开更多
基金supported by Natural Science Foundation of China(No.11771268)a center named Shanghai Research Center for Data Science and Decision Technology.
文摘In this paper,we study optimal model averaging estimators of regression coefficients in a multinomial logit model,which is commonly used in many scientific fields.A Kullback-Leibler(KL)loss-based weight choice criterion is developed to determine averaging weights.Under some regularity conditions,we prove that the resulting model averaging estimators are asymptotically optimal.When the true model is one of the candidate models,the averaged estimators are consistent.Simulation studies suggest the superiority of the proposed method over commonly used model selection criterions,model averaging methods,as well as some other related methods in terms of the KL loss and mean squared forecast error.Finally,the website phishing data is used to illustrate the proposed method.
文摘The multiple cause of death (MCOD) analysis is used to account for the full contribution of TB as a cause of death to South African mortality in 2008 that were coded using ICD10. Following a review of MCOD methods, a sufficient set of variables for use in MCOD and a new method of quantifying the severity of each cause of death are proposed. The results show that a total of 86,818 (14.3% of all deaths) were TB related, and within all deaths due to natural underlying causes, 86,373 (16.1%) were TB related. Furthermore, 42,581 (7.9%) were due to TB only, 6.0% had TB as an underlying cause along with other contributory causes and 2.0% had TB as a contributory cause. TB was mentioned as the underlying cause of death in 74,863 certificates or 13.9% of deaths due to natural underlying causes. Further analysis using multinomial baseline logit models, reveals that the relative odds of death in any demographic group compared with death in the baseline categories depend on the severity level of TB considered. It is proposed that the severity measure should be adopted when studying the contribution of all main causes of death to total mortality.
文摘This study applied multilevel modeling to investigate the impact of observed predictors and different levels or groups that households belong, on parents’ choice of discipline methods using data from 8156 households derived from a nationwide survey by the Ghana Statistical Service (GSS) in 2011. The aim of the study is to provide in-depth information on why parents choose particular discipline methods as corrective measures to reduce unwanted child behaviour in the present and to increase desirable ones in the future. The results of the study show that, religion and age-group of household heads have significant effect on household’s likelihood to choose physical discipline methods whereas the wealth index of a household and ethnicity of the household head, have significant effect on households’ likelihood to choose non-physical and psychological aggression methods. The results further show significant contextual effect on the differences in choices of parents at the household and regional levels. The choice of physical discipline methods by parents was consistent across households and regional levels unlike non-physical and psychological aggression methods whose application varied across the regions. Households in the Northern, Eastern and Volta regions mostly chose to apply physical discipline methods whereas in the Upper West, Western and Northern regions the most chosen discipline methods were non-physical discipline methods. Psychological aggression discipline methods were predominantly applied in the Upper East, Central and Northern regions of the country.
文摘A growing stream of study stresses the relevance of subjective elements in understanding the hierarchy of preferences that underpin individual travel behavior. The purpose of this study is to evaluate the impact of various factors on mode choice. To achieve this, a multinomial logit model (MNL) was used to analyze the relationships between mode choice and three classes of attributes;Combined Active and Latent, Active only and Latent only attributes. The data used are derived from surveys in the port city of Douala, Cameroon as a case study. Results stipulated that, the combined attributes model performed better than both active only attributes and latent only attributes models. Likewise, latent only attributes model performed better than active only attributes model. The advantage of modelling all three groups is for better selection of the most relevant attributes, and this is very relevant in understanding travel behavior of individuals and mode choice decisions.
