• 한국도로학회논문집
• /
• 제14권6호
• /
• pp.121-129
• /
• 2012

PURPOSES : Traffic accidents at intersections have been increased annually so that it is required to examine the causations to reduce the accidents. However, the current existing accident models were developed mainly with non-linear regression models such as Poisson methods. These non-linear regression methods lack to reveal complicated causations for traffic accidents, though they are right choices to study randomness and non-linearity of accidents. Therefore, to reveal the complicated causations of traffic accidents, this study used structural equation methods(SEM). METHODS : SEM used in this study is a statistical technique for estimating causal relations using a combination of statistical data and qualitative causal assumptions. SEM allow exploratory modeling, meaning they are suited to theory development. The method is tested against the obtained measurement data to determine how well the model fits the data. Among the strengths of SEM is the ability to construct latent variables: variables which are not measured directly, but are estimated in the model from several measured variables. This allows the modeler to explicitly capture the unreliability of measurement in the model, which allows the structural relations between latent variables to be accurately estimated. RESULTS : The study results showed that causal factors could be grouped into 3. Factor 1 includes traffic variables, and Factor 2 contains turning traffic variables. Factor 3 consists of other road element variables such as speed limits or signal cycles. CONCLUSIONS : Non-linear regression models can be used to develop accident predictions models. However, they lack to estimate causal factors, because they select only few significant variables to raise the accuracy of the model performance. Compared to the regressions, SEM has merits to estimate causal factors affecting accidents, because it allows the structural relations between latent variables. Therefore, this study used SEM to estimate causal factors affecting accident at urban signalized intersections.

• Advances in pediatric surgery
• /
• 제12권2호
• /
• pp.202-212
• /
• 2006

The prognostic factors for extrahepatic biliary atresia (EHBA) after Kasai portoenterostomy include the patient's age at portoenterostomy (age), size of bile duct in theporta hepatis (size), clearance of jaundice after operation (clearance) and the surgeon's experience. The aim of this study is to examine the most significant prognostic factor of EHBA after Kasai portoenterostomy. This retrospective study was done in 51 cases of EHBA that received Kasai portoenterostomy by one pediatric surgeon. For the statistical analysis, Kaplan-Meier method, Logrank test and Cox regression test were used. A p value of less than 0.05 was considered to be significant. Fifteen patients were regarded as dead in this study, including nine cases of liver transplantation. There was no significant difference of survival to age. The age is also not a significant risk factor for survival in this study (Cox Regression test; p = 0.63). There was no significant difference in survival in relation to the size of bile duct. However, bile duct size was a significant risk factor for survival (Cox Regression test; p = 0.002). There was a significant difference in relation to survival and clearance (Kaplan-Meier method; p = 0.02). The clearing was also a significant risk factor for survival (Cox Regression test; p = 0.001). The clearance of jaundice is the most significant prognostic factor of EHBA after Kasai portoenterostomy.

• 응용통계연구
• /
• 제33권5호
• /
• pp.569-578
• /
• 2020

분위수 회귀모형은 반응변수의 조건부 분위수 함수를 추정함으로써 반응변수와 예측변수의 관계에 대한 포괄적인 정보를 제공한다. 특히 커널 분위수 회귀모형은 비선형 관계식을 고려하기 위하여 양정치 커널함수(kernel function)에 의해 만들어지는 재생 커널 힐버트 공간(reproducing kernel Hilbert space)에서 비선형 조건부 분위수 함수를 추정한다. 그러나 KQR은 이차계획법으로 공식화되어 많은 계산비용을 필요로 하므로 컴퓨터 메모리 능력의 제한으로 대용량 자료의 분석은 불가능하다. 이러한 문제점을 해결하기 위하여 본 논문에서는 분할정복(divide and conquer) 알고리즘을 활용한 KQR 추정법(DC-KQR)을 제안한다. DC-KQR은 먼저 전체 훈련자료를 몇 개의 부분집합으로 무작위로 분할(divide)한 후, 각각의 부분집합에 대하여 KQR 분위수 함수를 추정하고 이들의 산술 평균을 이용하여 최종적인 추정량으로 통합(conquer)하는 기법이다. 본 논문에서는 모의실험과 실제자료 분석을 통해 제안한 DC-KQR의 효율적인 성능과 활용 가능성을 확인하였다.

• Imaging Science in Dentistry
• /
• 제49권1호
• /
• pp.19-26
• /
• 2019

Purpose: It has been proposed that using new prediction methods, such as neural networks based on dental data, could improve age estimation. This study aimed to assess the possibility of exploiting neural networks for estimating age by means of the pulp-to-tooth ratio in canines as a non-destructive, non-expensive, and accurate method. In addition, the predictive performance of neural networks was compared with that of a linear regression model. Materials and Methods: Three hundred subjects whose age ranged from 14 to 60 years and were well distributed among various age groups were included in the study. Two statistical software programs, SPSS 21 (IBM Corp., Armonk, NY, USA) and R, were used for statistical analyses. Results: The results indicated that the neural network model generally performed better than the regression model for estimation of age with pulp-to-tooth ratio data. The prediction errors of the developed neural network model were acceptable, with a root mean square error (RMSE) of 4.40 years and a mean absolute error (MAE) of 4.12 years for the unseen dataset. The prediction errors of the regression model were higher than those of the neural network, with an RMSE of 10.26 years and a MAE of 8.17 years for the test dataset. Conclusion: The neural network method showed relatively acceptable performance, with an MAE of 4.12 years. The application of neural networks creates new opportunities to obtain more accurate estimations of age in forensic research.

• 아태비즈니스연구
• /
• 제11권3호
• /
• pp.63-77
• /
• 2020

Purpose - The purpose of this study is to examine the effects of personal characteristics, loan characteristics, and interest rate characteristics of 2,653 borrowers on the delinquency possibility. In doing so, this study applies both multiple regression and logistic regression models to the data of credit unions in the city of Daegu. Design/Methodology/Approach - The major results of multiple regression analysis using SPSS are as follows. Findings - As for the results of testing the significance of the regression coefficients, it has been found that among the personal characteristics variables membership, credit rating, credit rating changes, and LTV have significant positive (+) effects on the delinquency possibility. Also it has been shown that among the loan characteristics variables loan amount, loan balance, total debt amount, collateral type, collateral amount, and repayment method have significant positive (+) effects on the delinquency possibility. Furthermore it has been found that among the interest rate characteristics variables both overdue interest rate and interest rate spread have positive (+) effects on the delinquency possibility. However, it has been shown that among the personal characteristics variables equity and membership do not have significant effects on the delinquency possibility, and that normal interest rate among the interest rate characteristics variables also do not have a significant effect on the delinquency possibility. Research Implications - By systematically analyzing the variables affecting delinquency possibility based on the results of this study, credit unions might get positive help in improving the system of managing receivables. Furthermore, the results of this study could be extended and applied to other types of financial institutions, so that financial institutions in general will also get some help to systematically manage the delinquency possibility.

• 한국IT서비스학회지
• /
• 제21권4호
• /
• pp.91-103
• /
• 2022

This study proposes a model for predicting contracted power using electric power data collected in real time from convenience stores nationwide. By optimizing the prediction model using machine learning, it will be possible to predict the contracted power required to renew the contract of the existing convenience store. Contracted power is predicted through the XGBoost regression model. For the learning of XGBoost model, the electric power data collected for 16 months through a real-time monitoring system for convenience stores nationwide were used. The hyperparameters of the XGBoost model were tuned using the GridesearchCV, and the main features of the prediction model were identified using the xgb.importance function. In addition, it was also confirmed whether the preprocessing method of missing values and outliers affects the prediction of reduced power. As a result of hyperparameter tuning, an optimal model with improved predictive performance was obtained. It was found that the features of power.2020.09, power.2021.02, area, and operating time had an effect on the prediction of contracted power. As a result of the analysis, it was found that the preprocessing policy of missing values and outliers did not affect the prediction result. The proposed XGBoost regression model showed high predictive performance for contract power. Even if the preprocessing method for missing values and outliers was changed, there was no significant difference in the prediction results through hyperparameters tuning.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권8호
• /
• pp.1976-1995
• /
• 2023

The anchor-free object detector CenterNet regards the object as a center point and predicts it based on the Gaussian circle region. For each object's center point, CenterNet directly regresses the width and height of the objects and finally gets the boundary range of the objects. However, the critical range of the object's center point can not be accurately limited by using the Gaussian circle region to constrain the prediction region, resulting in many low-quality centers' predicted values. In addition, because of the large difference between the width and height of different objects, directly regressing the width and height will make the model difficult to converge and lose the intrinsic relationship between them, thereby reducing the stability and consistency of accuracy. For these problems, we proposed a center point prediction method based on the Gaussian elliptic region and a size component regression method based on the small solution space. First, we constructed a Gaussian ellipse region that can accurately predict the object's center point. Second, we recode the width and height of the objects, which significantly reduces the regression solution space and improves the convergence speed of the model. Finally, we jointly decode the predicted components, enhancing the internal relationship between the size components and improving the accuracy consistency. Experiments show that when using CenterNet as the improved baseline and Hourglass-104 as the backbone, on the MS COCO dataset, our improved model achieved 44.7%, which is 2.6% higher than the baseline.

• 한국병원경영학회지
• /
• 제10권1호
• /
• pp.71-92
• /
• 2005

The purpose of this study is to analyze the efficiency in management of general hospitals and investigate the major factors on efficiency. Specifically, the management of each general hospital is evaluated by using Data Envelopment Analysis(DEA) technique which is a nonparametric statistical method for measurement of efficiency. Then, the influencing factors are investigated through analyses of Decision-Tree Model and Tobit Regression. The target hospitals were general hospitals in which bed sizes are between 200 and 500 among a total of 276 general hospitals. The main data of financial indicators were collected from 48 hospitals, and it was analyzed by using two statistical models. For Model I, three input and two output variables were used for efficiency evaluation. In particular, three input variables were the number of medical doctors, the number of paramedical personnel, and the bed size. And, two output variables were the numbers of inpatients and outpatients per year, adjusted by bed-size. The results of DEA analysis showed that only seven out of 48 hospitals(15%) turned out to be efficient. The decision-tree analysis also showed that there were six significant influencing factors for Model I. Six factors for Model I were Bed Occupancy Rate, Cost per Adjusted Inpatient, New Visit Ratio of Outpatients, Retired Ratio, Net Profit to Gross Revenues, Net Profit to Total Assets. In addition, the management efficiency of hospital is proved to increase as profit and patient-induced indicators increase and cost-related indicators decrease, by the Tobit regression model of independent variables derived from the decision-tree analysis. This study may be contributable to the development of analytic methodology regarding the efficiency of hospital management in that it suggests the synthetic measures by utilizing DEA model instead of suggesting simple ratio-analyzing results.

• 응용통계연구
• /
• 제27권4호
• /
• pp.629-643
• /
• 2014

회귀모형을 이용하여 자료를 분석하는 경우 이상점이나 영향점과 같은 특이값들의 유무를 검정하는 회귀진단기법은 모형의 적합성을 체크하기 위한 필수적인 도구로 잡은 지 오래이다. 이러한 점들이 존재 하는 경우 회귀분석의 결과가 왜곡되어 해석이 된다. Jang과 Anderson-Cook (2013)은 불꽃그림이란 이름을 붙인 그림도구를 발표하였는데 관측값에 부여된 가중치를 1에서 0으로 변화함에 따라 이상점이나 영향점이 회귀계수 및 잔차제곱합(SSE)에 어떠한 영향을 미치는지 3차원 그림에 추적곡선을 그려 보았을 뿐 아니라 쌍으로 대비시켜 봄으로써 분석의 시각적인 효과를 증대시켰다. 본 연구에서는 더 나아가 이러한 시도가 기존 방법과 어떤 차이점이 있는지 2013년에는 반영치 않은 통계량을 포함해서 더 많은 해석이 가능한지 혼합물 실험 계획을 통해 다양한 통계량의 민감도 분석을 실행하였다. 왜냐하면 작은 혼합물실험인 자료인 경우 더욱 세밀한 통계량에 대한 민감도 분석이 필요하기 때문이다.

• 산업경영시스템학회지
• /
• 제40권4호
• /
• pp.246-252
• /
• 2017

In the defense acquisition, a company's goal is to maximize profits, and the government's goal is to allocate budgets efficiently. Each year, the government estimates the ratio of indirect cost sector to defense companies, and estimates the ratio to be applied when calculating cost of the defense articles next year. The defense industry environment is changing rapidly, due to the increasing trend of defense acquisition budgets, the advancement of weapon systems, the effects of the 4th industrial revolution, and so on. As a result, the cost structure of defense companies is being diversifying. The purpose of this study is to find an alternative that can enhance the rationality of the current methodology for estimating the ratio of indirect cost sector of defense companies. To do this, we conducted data analysis using the R language on the cost data of defense companies over the past six years in the Defense Integrated Cost System. First, cluster analysis was conducted on the cost characteristics of defense companies. Then, we conducted a regression analysis of the relationship between direct and indirect costs for each cluster to see how much it reflects the cost structure of defense companies in direct labor cost-based indirect cost rate estimates. Lastly a new ratio prediction model based on regularized regression analysis was developed, applied to each cluster, and analyzed to compare performance with existing prediction models. According to the results of the study, it is necessary to estimate the indirect cost ratio based on the cost character group of defense companies, and the direct labor cost based indirect cost ratio estimation partially reflects the cost structure of defense companies. In addition, the current indirect cost ratio prediction method has a larger error than the new model.