• Journal of the Korean Statistical Society
• /
• v.36 no.3
• /
• pp.435-445
• /
• 2007

This paper addresses for the first time the use of complete data information criteria in unsupervised learning of the Naive-Bayes model. A Monte Carlo study sets a large experimental design to assess these criteria, unusual in the Bayesian network literature. The simulation results show that complete data information criteria underperforms the Bayesian information criterion (BIC) for these Bayesian networks.

• Communications for Statistical Applications and Methods
• /
• v.22 no.5
• /
• pp.507-518
• /
• 2015

It is not easy to select a linear mixed model since the main interest for model building could be different and the number of parameters in the model could not be clearly defined. In this paper, performance of conditional Akaike Information Criteria and its bias-corrected version are compared with marginal Bayesian and Akaike Information Criteria through a simulation study. The results from the simulation study indicate that bias-corrected conditional Akaike Information Criteria shows promising performance when candidate models exclude large models containing the true model, but bias-corrected one prefers over-parametrized models more intensively when a set of candidate models increases. Marginal Bayesian and Akaike Information Criteria also have some difficulty to select the true model when the design for random effects is nested.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.6
• /
• pp.1191-1197
• /
• 2009

Most Bayesian approaches to model selection of wavelet analysis have drawbacks that computational cost is expensive to obtain accuracy for the fitted unknown function. To overcome the drawback, this article introduces loss functions which are criteria for level dependent threshold selection in wavelet based Bayesian methods with arbitrary size and regular design points. We demonstrate the utility of these criteria by four test functions and real data.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.14
• /
• pp.5571-5575
• /
• 2014

Background: The ability to predict the survival time of breast cancer patients is important because of the potential high morbidity and mortality associated with the disease. To develop a predictive inference for determining the survival of breast cancer patients, we applied a novel Bayesian method. In this paper, we propose the development of a databased statistical probability model and application of the Bayesian method to predict future survival times for White Hispanic female breast cancer patients, diagnosed in the US during 1973-2009. Materials and Methods: A stratified random sample of White Hispanic female patient survival data was selected from the Surveillance Epidemiology and End Results (SEER) database to derive statistical probability models. Four were considered to identify the best-fit model. We used three standard model-building criteria, which included Akaike Information Criteria (AIC), Bayesian Information Criteria (BIC), and Deviance Information Criteria (DIC) to measure the goodness of fit. Furthermore, the Bayesian method was used to derive future survival inferences for survival times. Results: The highest number of White Hispanic female breast cancer patients in this sample was from New Mexico and the lowest from Hawaii. The mean (SD) age at diagnosis (years) was 58.2 (14.2). The mean (SD) of survival time (months) for White Hispanic females was 72.7 (32.2). We found that the exponentiated Weibull model best fit the survival times compared to other widely known statistical probability models. The predictive inference for future survival times is presented using the Bayesian method. Conclusions: The findings are significant for treatment planning and health-care cost allocation. They should also contribute to further research on breast cancer survival issues.

• Advances in concrete construction
• /
• v.1 no.1
• /
• pp.85-102
• /
• 2013

Prior concrete strength distributions can be updated by using direct information from test results as well as by taking into account indirect information due to conformity control. Due to the filtering effect of conformity control, the distribution of the material property in the accepted inspected lots will have lower fraction defectives in comparison to the distribution of the entire production (before or without inspection). A methodology is presented to quantify this influence in a Bayesian framework based on prior knowledge with respect to the hyperparameters of concrete strength distributions. An algorithm is presented in order to update prior distributions through numerical integration, taking into account the operating characteristic of the applied conformity criteria, calculated based on Monte Carlo simulations. Different examples are given to derive suitable hyperparameters for incoming strength distributions of concrete offered for conformity assessment, using updated available prior information, maximum-likelihood estimators or a bootstrap procedure. Furthermore, the updating procedure based on direct as well as indirect information obtained by conformity assessment is illustrated and used to quantify the filtering effect of conformity criteria on concrete strength distributions in case of a specific set of conformity criteria.

• The Korean Journal of Applied Statistics
• /
• v.22 no.1
• /
• pp.197-204
• /
• 2009

In the model selection problem, the main objective is to choose the true model from a manageable set of candidate models. An information criterion gauges the validity of a statistical model and judges the balance between goodness-of-fit and parsimony; "how well observed values ran approximate to the true values" and "how much information can be explained by the lower dimensional model" In this study, we introduce some information criteria modified from the Akaike Information Criterion (AIC) and the Bayesian Information Criterion(BIC). The information criteria considered in this study are compared via simulation studies and real application.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.2
• /
• pp.663-669
• /
• 2014

Background: With recent progress in health science administration, a huge amount of data has been collected from thousands of subjects. Statistical and computational techniques are very necessary to understand such data and to make valid scientific conclusions. The purpose of this paper was to develop a statistical probability model and to predict future survival times for male breast cancer patients who were diagnosed in the USA during 1973-2009. Materials and Methods: A random sample of 500 male patients was selected from the Surveillance Epidemiology and End Results (SEER) database. The survival times for the male patients were used to derive the statistical probability model. To measure the goodness of fit tests, the model building criterions: Akaike Information Criteria (AIC), Bayesian Information Criteria (BIC), and Deviance Information Criteria (DIC) were employed. A novel Bayesian method was used to derive the posterior density function for the parameters and the predictive inference for future survival times from the exponentiated Weibull model, assuming that the observed breast cancer survival data follow such type of model. The Markov chain Monte Carlo method was used to determine the inference for the parameters. Results: The summary results of certain demographic and socio-economic variables are reported. It was found that the exponentiated Weibull model fits the male survival data. Statistical inferences of the posterior parameters are presented. Mean predictive survival times, 95% predictive intervals, predictive skewness and kurtosis were obtained. Conclusions: The findings will hopefully be useful in treatment planning, healthcare resource allocation, and may motivate future research on breast cancer related survival issues.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.4
• /
• pp.260-266
• /
• 2017

The global economy, including Korea, has continuously searched for various market-friendly policies and new economic systems in pursuit of the forth industrial revolution. As a result, economic markets have grown, and factors affecting markets have diversified. Therefore, as for many company's decision makers, it has become an important issue to analyze and forecast markets accurately and effectively for rapid and appropriate decision making. In this study, we aim to improve the accuracy and validity of forecast models by applying extended information criteria in existing restricted information criteria to determine optimized modeling for the accurate analysis and prediction of complex market environments. In order to verify the practical use of the extended information criteria adopted in this study, we compare this study employing KOSPI data with previous studies. Experimental results show that applying extended information criteria is more accurate than using the existing information criteria.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.19
• /
• pp.8371-8376
• /
• 2014

Background: The use of statistical methods has become an imperative tool in breast cancer survival data analysis. The purpose of this study was to develop the best statistical probability model using the Bayesian method to predict future survival times for the black non-Hispanic female breast cancer patients diagnosed during 1973-2009 in the U.S. Materials and Methods: We used a stratified random sample of black non-Hispanic female breast cancer patient data from the Surveillance Epidemiology and End Results (SEER) database. Survival analysis was performed using Kaplan-Meier and Cox proportional regression methods. Four advanced types of statistical models, Exponentiated Exponential (EE), Beta Generalized Exponential (BGE), Exponentiated Weibull (EW), and Beta Inverse Weibull (BIW) were utilized for data analysis. The statistical model building criteria, Akaike Information Criteria (AIC), Bayesian Information Criteria (BIC), and Deviance Information Criteria (DIC) were used to measure the goodness of fit tests. Furthermore, we used the Bayesian approach to obtain the predictive survival inferences from the best-fit data based on the exponentiated Weibull model. Results: We identified the highest number of black non-Hispanic female breast cancer patients in Michigan and the lowest in Hawaii. The mean (SD), of age at diagnosis (years) was 58.3 (14.43). The mean (SD), of survival time (months) for black non-Hispanic females was 66.8 (30.20). Non-Hispanic blacks had a significantly increased risk of death compared to Black Hispanics (Hazard ratio: 1.96, 95%CI: 1.51-2.54). Compared to other statistical probability models, we found that the exponentiated Weibull model better fits for the survival times. By making use of the Bayesian method predictive inferences for future survival times were obtained. Conclusions: These findings will be of great significance in determining appropriate treatment plans and health-care cost allocation. Furthermore, the same approach should contribute to build future predictive models for any health related diseases.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.6
• /
• pp.1531-1541
• /
• 2013

This paper proposes a new method of VaR forecasting using the conditional autoregressive VaR (CAViaR) models and information criteria. Instead of using a single CAViaR model, we propose to utilize several candidate CAViaR models during a forecasting period. By adopting the Akaike and Bayesian information criteria for quantile regression, we can update not only parameter estimates but also the CAViaR specifications. We also propose extended CAViaR models with a constant location parameter. An empirical study is provided to examine the performance of the proposed method. The results suggest that our method shows more stable performance than those using a single specification.