• Communications for Statistical Applications and Methods
• /
• v.9 no.2
• /
• pp.515-520
• /
• 2002

In general, we use the hierarchical Poisson-gamma model for the Poisson data in generalized linear model. Time effect will be emphasized for the analysis of the observed data to be collected annually for the time period. An extended model with time effect for estimating the effect is proposed. In particularly, we discuss the Quasi likelihood function which is used to numerical approximation for the likelihood function of the parameter.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.5 no.1
• /
• pp.85-105
• /
• 1981

Information is an essential factor leading the rapid progress which is one of the distinguished characteristics in modem society. As more information is required and as more is supplied by individuals, governmental units, businesses, and educational institutions, the greater will be the requirement for efficient methods of communication. One possibility for improving the information dissemination process is to use computers. The capabilities of such machine are beginning to be used in the process of Information storage, retrieval and dissemination. An important problems, that must be carefully examined is whether one technique for information retrieval is better for worse than another. This paper examines problem of how to evaluate an information retrieval system. One specific approach is a cost accounting model for use in studying how to minimize the cost of operating a mechanized retrieval system. Through the use of cost analysis, the model provides a method for comparative evaluation between systems. The general cost accounting model of the literature retrieval system being designed by this study are given below. 1. The total cost accounting model of the literature retrieval system. The total cost of the literature retrieval system = (the cost per unit of user time X the amount of user time) + ( the cost per unit of system time X the amount of system time) 2. System cost accounting model system cost = (the pre-search system cost per unit of time X time) + (the search system cost per unit of time X time) + (the post search system cost per unit of time X time) 1) Pre-search system cost per unit of time = cost of channel per unit time + cost of central processing unit per unit time + cost of storage per unit time 2) Search system cost per unit of time = comparison cost + document representation cost. 3) Post-search system cost per unit of time. = cost of channel per unit time + cost of central processing unit per unit time + cost of storage per unit time 3. User cost accounting model Total user cost = [pre-search user cost per unit of time X (time + additional time) ] + [search user cost per unit of time X (time + additional time) ] + [post-search user cost per unit of time X (time + additional time) ].

• International Journal of Highway Engineering
• /
• v.9 no.4
• /
• pp.21-32
• /
• 2007

The linear traffic model(Vertical queueing model) that is adopted widely in traffic flow estimation assumes that all vehicles have the identical motion before joining a queue at the stop-line. Thus, a queue is supposed to form vertically not horizontally. Due to the simplicity of this model, the departure time of the leading vehicle is assumed to coincide with the start of effective green time. Thus, the delay estimates given by the Vertical queueing model is not always realistic. This paper explores a microscopic traffic model(a Kinematic Car-following model at Signalised intersections: a KCS traffic model) based on the one dimensional Kinematic equations in physics. A comparative evaluation in delay and sensitivity of delay difference between the KCS traffic model and the previously known Vertical queueing model is presented. The results show that the delay estimate in the Vertical queueing model is always greater than or equal to the KCS traffic model; however, the sensitivity of delay in the KCS traffic model is greater than the Vertical queueing model.

• Communications for Statistical Applications and Methods
• /
• v.23 no.6
• /
• pp.467-478
• /
• 2016

The accelerated failure time model or accelerated life model relates the logarithm of the failure time linearly to the covariates. The parameters in the model provides a direct interpretation. In this paper, we review some newly developed practically useful estimation and inference methods for the model in the analysis of right censored data.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.4
• /
• pp.1147-1157
• /
• 2005

Bonus-Malus system is generally constructed based on claim frequency and Bayesian credibility model is used to represent claim frequency distribution. However, there is a problem with traditionally used credibility model for the purpose of constructing bonus-malus system. In traditional Bonus-Malus system adopted credibility model, individual estimates of premium rates for insureds are determined based solely on the total number of claim frequency without considering when those claims occurred. In this paper, a new model which is a modification of structural time series model applicable to counting time series data are suggested. Based on the suggested model relatively higher premium rates are charged to insured with more claim records.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.6
• /
• pp.3273-3308
• /
• 2017

Building appropriate trust evaluation models is an important research issue for security guarantee in social networks. Most of the existing works usually consider the trust values at the current time slot, and model trust as the stochastic variable. However, in fact, trust evolves over time, and trust is a stochastic process. In this paper, we propose a novel time-variant stochastic trust evaluation (TSTE) model, which models trust over time and captures trust evolution by a stochastic process. Based on the proposed model, we derive the time-variant bound of untrustworthy probability, which provides stochastic trust guarantee. On one hand, the time-variant trust level of each node can be measured by our model. Meanwhile, by tolerating nodes with relatively poor performance, our model can effectively improve the node resource utilization rate. Numerical simulations are conducted to verify the accuracy and consistency of the analytical bounds on distinguishing misbehaved nodes from normal ones. Moreover, simulation results on social network dataset show the tradeoff between trust level and resource utilization rate, and verify that the successful transmission rate can be improved by our model.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.37 no.4
• /
• pp.37-50
• /
• 2012

The existing DEA models have been devoted to evaluate relative efficiency of DMUs based on multiple input and output factors of a same period. However, a certain kind of lead time can be required to produce outputs using inputs in an organization. R&D evaluation is a typical area with this kinds of time lag. Thus, the purpose of this paper is to develop a new DEA model to deal with time lag effect in performance evaluation. The proposed model is to find relative efficiency of each DMU for each period considering the time lag effect. A case example using a real data set is also given to show the usage or implication of the suggested model. The results are compared with the ones of the CCR model and the multi-periods input model.

• ETRI Journal
• /
• v.32 no.2
• /
• pp.302-311
• /
• 2010

In this study, we analyzed consumers' choices of various broadcasting and telecommunication convergence services and time consumption for chosen services by using survey data. A multivariate probit model was used to model consumers' choices of various broadcasting and telecommunication convergence services, and an ordered probit model was used to model consumers' time consumption for chosen services. Factors affecting consumers' choices and time-consumption behavior were identified, and simulation results of market competition and substitution were obtained. Based on these results, it was found that for the time being, consumers are highly locked into existing broadcasting services and are likely to become more price-sensitive to the new broadcasting and telecommunication convergence services. Also, the ways in which individual characteristics affect choices and time consumption were found to be very diverse service by service.

• Communications for Statistical Applications and Methods
• /
• v.18 no.3
• /
• pp.319-331
• /
• 2011

In this paper, we propose automatic procedures for the model selection of various univariate time series data. Automatic model selection is important, especially in data mining with large number of time series, for example, the number (in thousands) of signals accessing a web server during a specific time period. Several methods have been proposed for automatic model selection of time series. However, most existing methods focus on linear time series models such as exponential smoothing and autoregressive integrated moving average(ARIMA) models. The key feature that distinguishes the proposed procedures from previous approaches is that the former can be used for both linear time series models and nonlinear time series models such as threshold autoregressive(TAR) models and autoregressive moving average-generalized autoregressive conditional heteroscedasticity(ARMA-GARCH) models. The proposed methods select a model from among the various models in the prediction error sense. We also provide an R package autots that implements the proposed automatic model selection procedures. In this paper, we illustrate these algorithms with the artificial and real data, and describe the implementation of the autots package for R.

• Journal of Positioning, Navigation, and Timing
• /
• v.13 no.1
• /
• pp.111-115
• /
• 2024

This study builds a machine learning model optimized for clocks among various techniques in the field of artificial intelligence and applies it to clock stabilization or synchronization technology based on atomic clock noise characteristics. In addition, the possibility of providing stable source clock data is confirmed through the characteristics of machine learning predicted values during holdover of atomic clocks. The proposed machine learning model is evaluated by comparing its performance with the AutoRegressive Integrated Moving Average (ARIMA) model, an existing statistical clock prediction model. From the results of the analysis, the prediction model proposed in this study (MSE: 9.47476) has a lower MSE value than the ARIMA model (MSE: 221.2622), which means that it provides more accurate predictions. The prediction accuracy is based on understanding the complex nature of data that changes over time and how well the model reflects this. The application of a machine learning prediction model can be seen as a way to overcome the limitations of the statistical-based ARIMA model in time series prediction and achieve improved prediction performance.