• Journal of the Korean Operations Research and Management Science Society
• /
• v.37 no.1
• /
• pp.29-43
• /
• 2012

We design a DEA-AR model using multiple regression analysis with new methods which limit weights. When there are multiple input and single output variables, our model can be used, and the weights of input variables use the regression coefficient and coefficient of determination. To verify the effectiveness of the new model, we evaluate the efficiency of the Regional Corporations in Korea. Accordance with statistical analysis, it proved that there is no difference between the efficiency value of the DEA-AR using AHP and our DEA-AR model. Our model can be applied to a lot of research by substituting DEA-AR model relying on AHP in the future.

• Journal of the Korean Regional Science Association
• /
• v.14 no.2
• /
• pp.35-50
• /
• 1998

This paper examines the determinants of imbalanced regional development in Korea during the period of 1985-1995. The review of previous analytical techniques have been used to analyze the determinants of disparities in regional development of disparities in regional development, but few has applied the regression technique which reduces a bias due to heterogeneity across region. The results of the study show that Kmenta model with per capita GRDP as dependent variable can reduce the heterogeneity bias in regional development and can minimize the statical errors in estimation and interpretation of the coefficients of the explanatory variables. According to the results of Kmenta model, urban infrastructure such as roads, information and communication facilities are major causes of regional disparity over the period of 1985-1995. The results of the study also indicate that local government should devote their policy efforts to identify and utilize the unique soci-economic characteristics of each locality in the process of regional development.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.57 no.2
• /
• pp.37-45
• /
• 2015

Typically, water quality sampling takes place intermittently since sample collection and following analysis requires substantial cost and efforts. Therefore regression models (or rating curves) are often used to interpolate water quality data. LOADEST has nine regression models to estimate water quality data, and one regression model needs to be selected automatically or manually. The nine regression models in LOADEST and auto-selection by LOADEST were evaluated in the study. Suspended solids data were collected from forty-nine stations from the Water Information System of the Ministry of Environment. Suspended solid data from each station was divided into two groups for calibration and validation. Nash-Stucliffe efficiency (NSE) and coefficient of determination ($R_2$) were used to evaluate estimated suspended solid loads. The regression models numbered 1 and 3 in LOADEST provided higher NSE and $R_2$, compared to the other regression models. The regression modes numbered 2, 5, 6, 8, and 9 in LOADEST provided low NSE. In addition, the regression model selected by LOADEST did not necessarily provide better suspended solid estimations than the other regression models did.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.36 no.3
• /
• pp.407-416
• /
• 2016

The purpose of this study is to develop regional regression models to estimate mean low flow for ungauged basins. The unregulated streamflow data observed at 12 multipurpose dams and 4 irrigation dams were analyzed for determining mean low flows. Various types of regression models were developed using the relationship between mean low flows and various sets of watershed characteristics such as drainage area, average slope, drainage density, mean annual precipitation, runoff curve number. The performance of each regression model for estimating mean low flows was assessed by comparison with the results obtained from the observed data. It was found that a regional regression model explained by drainage area, the mean annual precipitation, and runoff curve number showed the best performance. The regression model presented in this study also gives better estimates of mean low flow than the estimates by the drainage-area ratio method and the previous regression model.

• Health Policy and Management
• /
• v.26 no.4
• /
• pp.271-278
• /
• 2016

Background: This study purposed to analyze the relationship between regional obesity rates and regional variables. Methods: Data was collected from the Korean Statistical Information Service (KOSIS) and Community Health Survey in 2012. The units of analysis were administrative districts such as city, county, and district. The dependent variable was the age-sex adjusted regional obesity rates. The independent variables were selected to represent four aspects of regions: health behaviour factor, psychological factor, socio-economic factor, and physical environment factor. Along with the traditional ordinary least square (OLS) regression analysis model, this study applied geographically weighted regression (GWR) analysis to calculate the regression coefficients for each region. Results: The OLS results showed that there were significant differences in regional obesity rates in high-risk drinking, walking, depression, and financial independence. The GWR results showed that the size of regression coefficients in independent variables was differed by regions. Conclusion: Our results can help in providing useful information for health policy makers. Regional characteristics should be considered when allocating health resources and developing health-related programs.

• Journal of Korea Water Resources Association
• /
• v.31 no.1
• /
• pp.59-69
• /
• 1998

A regression model with basin physiographic characteristics as independent variables was calibrated for regional flood frequency analysis. In case that high correlations existing among the independent variables the ridge regression has been known to have capability of overcoming the problems of multicollinearity. To optimize the ridge regression model the cost function including regularization parameter must be minimized. In this research the genetic algorithm was applied on this optimization problem. The genetic algorithm is a stochastic search method that mimic the metaphor of natural biological heredity. Using this method the regression model could have optimized and stable weights of variables.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.170-170
• /
• 2020

Due to the development of technologies, complex computation of huge data set is possible with a prevalent personal computer. Therefore, machine learning methods have been widely applied in the hydrologic field such as regression-based regional frequency analysis (RFA). The main purpose of this study is to compare two frameworks of RFA based on the artificial neural network (ANN) models: quantile regression technique (QRT-ANN) and parameter regression technique (PRT-ANN). As an output layer of the ANN model, the QRT-ANN predicts quantiles for various return periods whereas the PRT-ANN provides prediction of three parameters for the generalized extreme value distribution. Rainfall gauging sites where record length is more than 20 years were selected and their annual maximum rainfalls and various hydro-meteorological variables were used as an input layer of the ANN model. While employing the ANN model, 70% and 30% of gauging sites were used as training set and testing set, respectively. For each technique, ANN model structure such as number of hidden layers and nodes was determined by a leave-one-out validation with calculating root mean square error (RMSE). To assess the performances of two frameworks, RMSEs of quantile predicted by the QRT-ANN are compared to those of the PRT-ANN.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.36 no.3
• /
• pp.427-437
• /
• 2016

The objective of this study is to develop the regional regression models based on the physiographical and climatological characteristics for estimating flow duration curve (FDC) in ungauged bsisns. To this end, the lower sections with duration from 185 to 355 days of FDCs were constructed from the 16 gauged streamflow data, which were fitted to the two-parameter logarithmic type regression equation. Then, the parameters of the equation were regionalized using the basin characteristics such as basin area, basin slope, drainage density, mean annual precipitation, mean annual streamflow, runoff curve number in order that the proposed regression model can be used for ungauged basin. From the comparison of the estimated by the regional regression model with the observed ones, the model with the combination of basin area, runoff curve number, mean annual precipitation showed the best performance.

• Journal of Urban Science
• /
• v.9 no.1
• /
• pp.41-49
• /
• 2020

This study analyzed the factors affecting economic growth using multiple regression model and Geographically Weighted Regression in consideration of population, industry and employment, housing and political characteristics on economic growth by region. The analysis results are summarized as follows. First, the total employment growth rate, manufacturing employment growth rate, local election turnout and the level of party consensus between the central and local governments are having a positive impact on regional economic growth. Second, according to the GWR analysis, the population has a positive impact on economic growth in the southern region of Korea, and the increase in the total number of employees has a positive impact on the southern region of Gyeonggi Province, Gangwon Province, North Chungcheong Province and North Gyeongsang Province. Finally, the voter turnout of urbanites is positively affecting economic growth in South Chungcheong Province, Gangwon Province and the southern coast, while North Jeolla and South Jeolla provinces have a positive impact on economic growth as the parties of the central and local governments are equal. The results of this study may suggest the role of local government for regional economic development.

• Journal of Korean Society of Forest Science
• /
• v.83 no.4
• /
• pp.521-530
• /
• 1994

Voume functions, which are usually expressed by the function of dbh and height, are estimated commonly through the regression analysis with the highest statistical accuracy considered. In Korea, general volume functions for each tree species were prepared by means of the regression analysis with the exponential function ($V=aD^bH^c$) having the dbh(D) and height(H) as independent variables. In this study, regional stem curve functions for the Pinus densiflora in Kangwon-province were derived and a regional volume function model, in which the stem volume can be directly estimated through the rotational integral of the regional stem curve functions, was prepared. The regional volume estimated by the prepared model was more accurate than the volume by the general volume table for the Pinus densiflora in Kangwon-province. Additionary, the form of stem curves derived by the regional stem curve functions showed difference from each other. The stem in Youngwol and Wonju taper down more fast in upper part than that in other regions. These various stem forms also led to the regional difference in volume estimates.