• Journal of the Korean Geographical Society
• /
• v.51 no.2
• /
• pp.219-233
• /
• 2016

In social science fields, statistical models are used almost exclusively for causal explanation, and explanatory modeling has been a mainstream until now. In contrast, predictive modeling has been rare in the fields. Hence, we focus on constructing the predictive non-parametric model, instead of the explanatory model. Gangnam-gu, Seoul was chosen as a study area and we collected single-family house sales data sold between 2011 and 2014. We applied non-parametric models proposed in machine learning area including generalized additive model(GAM), random forest, multivariate adaptive regression splines(MARS) and support vector machines(SVM). Models developed recently such as MARS and SVM were found to be superior in predictive power for house price estimation. Finally, spatial autocorrelation was accounted for in the non-parametric models additionally, and the result showed that their predictive power was enhanced further. We hope that this study will prompt methodology for property price estimation to be extended from traditional parametric models into non-parametric ones.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.4
• /
• pp.183-194
• /
• 2021

In this study, the conversion factor for 2020 is estimated based on an outpatient and inpatient conversion factor separation model developed from SGR and AR by using actual medical expense data. In addition, a policy plan is proposed to calculate the values of single and multiple conversion factors for each type of medical expense, and to effectively use the conversion factor separation model as one of the means to establish a medical delivery system. The major results are as follows. First, at r=0.1, the rate of adjustment in the hospital single conversion index in 2020 was 2.0%, and the outpatient and hospitalization conversion rates for hospitals were 2.2% and 2.3%, respectively. In addition, a combination of outpatient and inpatient conversion factors can be used for the adjustment. Second, as a measure to establish a medical delivery system, instead of adjusting the addition rate, a method of interlocking the addition rate and the conversion factor is proposed. Third, it is necessary to develop a model that enables target management of volumes, in addition to the outpatient conversion factor, the inpatient conversion factor, and the adjustment coefficient.

• The Korean Journal of Applied Statistics
• /
• v.18 no.1
• /
• pp.43-56
• /
• 2005

In this study, a robust estimation method for the first-order autocorrelation coefficient in the time series model following AR(l) process with additive outlier(AO) is investigated. We propose the L-type trimmed least squares estimation method using the preliminary estimator (PE) suggested by Rupport and Carroll (1980) in multiple regression model. In addition, using Mallows' weight function in order to down-weight the outlier of X-axis, the bounded-influence PE (BIPE) estimator is obtained and the mean squared error (MSE) performance of various estimators for autocorrelation coefficient are compared using Monte Carlo experiments. From the results of Monte-Carlo study, the efficiency of BIPE(LAD) estimator using the generalized-LAD to preliminary estimator performs well relative to other estimators.

• The Korean Journal of Applied Statistics
• /
• v.31 no.6
• /
• pp.733-749
• /
• 2018

Sample selection arises as a result of the partial observability of the outcome of interest in a study. Heckman introduced a sample selection model to analyze such data and proposed a full maximum likelihood estimation method under the assumption of normality. Recently sample selection models for binomial and Poisson response variables have been proposed. Based on the theory of symmetry-modulated distribution, we extend these to a model for overdispersed count data. This type of data with no sample selection is often modeled using negative binomial distribution. Hence we propose a sample selection model for overdispersed count data using the negative binomial distribution. A real data application is employed. Simulation studies reveal that our estimation method based on profile log-likelihood is stable.

• Journal of Korean Society of Transportation
• /
• v.29 no.5
• /
• pp.7-17
• /
• 2011

Traffic problems caused by drunk drivers have been steadily raised from the past. Even though the previous researches have focused on the development of countermeasures for preventing drunk driving, the number of drivers violating the DUI (Driving-Under-Influence) regulation is still increasing. Many studies seek countermeasures for preventing drunk driving by comparing the differences between general and drunk drivers. However, few researches have investigated focusing only on the characteristics of drunk drivers. It is well known that characteristics of general drivers are different from those of drunk drivers, and also habitual drunk drivers have different characteristics from non-habitual drunk drivers. Motivated by this fact, only the drivers who have violated DUI regulation are considered in the analysis. This study primarily aims to provide alternative solutions for reducing habitual drunk drivers who are highly inclined to do drunk driving repeatedly. For the analysis, various types of variables potentially effecting drunk driving behavior were investigated, and then truncated count data models were developed to analyze the effects of the variables selected on drunk driving. The results showed that 1) a truncated negative binomial model is better fitted to the data; and 2) five variables including experiential learning, the lack of self-control, self-reflection, the fear of crackdown, and the level of dependence on vehicles were found to be statistically significant.

• Journal of Technology Innovation
• /
• v.14 no.3
• /
• pp.23-42
• /
• 2006

This study investigates the determinants of output of a manufacturing firm's innovative activity (the number of patent applications) and spillover effects using a count data model. This paper attempted a negative binomial distribution In order to take into account unobserved heterogeneity. The results of our study suggested that Firm size, R&D intensity, technical network activity, and online business performance have significantly positive effects in the Korean manufacturing industry. Moreover, there are significantly positive spillover effects in the same industry sector.

• Journal of the Korean Regional Science Association
• /
• v.31 no.4
• /
• pp.107-128
• /
• 2015

Technology innovation is a competitive weapon of sustainable economic growth at the urban and regional level and the growth of firms. In this study, we empirically investigate the effects of collaborative R&D activity on product innovative outputs and process innovative outputs in manufacturing firms in Korea. We analyze the links between collaborative R&D activity and two types of innovative outputs using an alternative negative binomial regression model. The major finding is that collaborative R&D activity has significant positive effects on both product and process innovation. The results also identify a positive link between all types of innovative outputs and other R&D activities including internal R&D activity, patent activity, external technology and capital goods acquisitions. To induce corporate growth that enhances the productivity of individual firms and produces prolonged economic growth, policy makers should place greater emphasis on creating effective arrangements to promote establishing collaborative R&D strategies for manufacturing firms.

• Journal of Korean Society of Transportation
• /
• v.34 no.6
• /
• pp.489-498
• /
• 2016

The objective of this study is to develop road traffic accident model involving child pedestrian especially at school zones and its surrounding area. The analysis is based upon traffic accident data collected near sixty elementary schools in City of Cheongju during 2012 and 2014. This study results in two statistical models ; one is to predict the number of road traffic accidents involving children, and the other is to predict EPDO(Equivalent Prperty Damage Only). These models are represented as Poisson models. which are statistically significant with the likelihood ratios of 0.533 and 0.273. The common explanatory variables of these models are the ratio of road section with more than 4 lanes, the number of entrance and exit, the number of signalized crosswalk in school zone, the number of school zone signage including road surface marking, and the number of speed limit signs. The specific variables are the length of road stretch in school zone, the number of reflector mirrors, and the number of signalized crosswalk outside school zone. It is concluded that these types of road safety facilities can reduce the number of traffic accidents involving children at school zones and its surrounding area.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.36 no.2
• /
• pp.333-340
• /
• 2016

The aim of this study is to specifically understand the characteristics of the crime by empirical analysis for the determining factors that affect determining the crime through the space syntax in Busan. In this study, poisson regression and negative binomial regression were used for accurate analysis. 8 variables that were significant of the total 13 variables. The summary if this study based on the results is as follow. Statistically significant variables are female ratio, over 65 population ratio, administration are and commercial area ratio in characteristics. And the more CCTVs a region has, the lower crime rate it shows. As a results of examing whether space syntax variables can predict crime occurrence places. Space with low connectivity come to be a crime causal factor because they have few other related spaces and thereby have low possibility of sudden appearance of interrupters, which results in low surveillance levels of foot passengers. It will provide the basic data that can contribute to urban planning and implementation of crime prevention aspects.

• Management & Information Systems Review
• /
• v.33 no.2
• /
• pp.301-315
• /
• 2014

The purpose of this study is to estimate the economic value of the Kumjung Mountain, using a Individual Travel Cost Model(ITCM). This paper compares Poisson and negative binomial count data models to measure the tourism demands. Interviewers were instructed to interview only individuals. So the sample was taken in 700. A dependent variable that is defined on the non-negative integers and subject to sampling truncation is the result of a truncated count data process. The results suggest that the truncated negative binomial model is improved overdispersion problem and more preferred than the other models in the study. This study emphasizes in particular 'travel cost' that is not only monetary cost but also including opportunity cost of 'travel time'. According to the truncated negative binomial model, estimates the Consumer Surplus(CS) values per trip of about 60,669 Korean won and the total economic value was estimated to be 252,383 Korean won.