• Journal of the Korean Data and Information Science Society
• /
• v.21 no.5
• /
• pp.901-908
• /
• 2010

A geometrical description method is proposed to represent the process of the forward selection and backward elimination methods among many variable selection methods for multiple regression models. This graphical method shows the process of the forward selection and backward elimination on the first and second quadrants, respectively, of half circle with a unit radius. At each step, the SSR is represented by the norm of vector and the extra SSR or partial determinant coefficient is represented by the angle between two vectors. Some lines are dotted when the partial F test results are statistically significant, so that statistical analysis could be explored. This geometrical description can be obtained the final regression models based on the forward selection and backward elimination methods. And the goodness-of-fit for the model could be explored.

• The Korean Journal of Applied Statistics
• /
• v.22 no.2
• /
• pp.297-307
• /
• 2009

Korean household debt is one of the main sources of the current financial crisis. This paper studies the impact of household members' attributes such as a type of housing(self-own or rent), education, age, average monthly income of the head of household, and the area of residence, on the stress level of the household members due to household debt. We analyze a real data set collected by KB Kookmin Bank in 2004. We consider low and high stress level as a binary response variable and use a logistic regression model with the attributes of household members as explanatory variables. A simple but well-fitting model is selected by backward elimination method based on the likelihood statistic for goodness-of-fit test, and the impact of the attributes on the stress level is studied from parameter estimates of the selected model. We also perform the similar analysis on a binary response variable which distinguishes households with no debt from the rest. From the analysis, the stress level tends to be low for households with self-own houses, high average monthly income, low education level, and young members.

• Journal of the Korean Data and Information Science Society
• /
• v.23 no.3
• /
• pp.457-465
• /
• 2012

The purpose of this study is to analyze the association between indicators and to find statistical models based on important indicators at 'College Notifier' in Korea Council for University Education. First, Pearson correlation coefficients are used to find statistically significant correlations. By variable selection method, the important indicators are selected and their coefficients are estimated. As variable selection method, backward and stepwise methods are employed.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.04a
• /
• pp.683-685
• /
• 2000

이 논문에서는 새로운 중앙 소거 방식과 행렬 이분법의 개념을 시스톨릭한 구조 위에 결합시켜{{{{ { O}^{ } }}}}({{{{ { N}^{ 3} }}}})의 문제를 해결한다. 새로운 중앙 소거 방식은 주어진 시스톨릭한 구조의 병렬성을 최대한 증가시키는 것을 가능하게 해주며, 행렬 이분법은 기존의 가우스 소거 방식 상에서 나타나는 {{{{ { O}^{ } }}}}({{{{ { N}^{ 2} }}}})의 복잡도를 요구하는 후진 대입을 효과적으로 제거 시켜준다. 새로운 소거 방법은 독립적인 선형방정식으로 이루어진 시스템의 차수를 N이라 할 때 2N(N+1)의 저장 공간과 4N+2log2N-4의 시간 복잡도를 갖는다. 제안 한 새로운 소거 방식은 단순한 구조와 연결 방식을 가진 그물 구조의 시스톨릭 병렬 시스템에 적용되기에 충분히 적합한 단순한 알고리즘을 사용하면서도 이전의 방법과 동일한 구조의 저정공간을 요구하고 동시에 훨씬 우수한 시간 성능을 나타내는 것이 가능하다.

• The Korean Journal of Applied Statistics
• /
• v.18 no.2
• /
• pp.435-443
• /
• 2005

Coefficients of determination in logistic regression analysis are defined as various statistics, and their values are relatively smaller than those for linear regression model. These coefficients of determination are not generally used to evaluate and diagnose logistic regression model. Liao and McGee (2003) proposed two adjusted coefficients of determination which are robust at the addition of inappropriate predictors and the variation of sample size. In this work, these adjusted coefficients of determination are applied to variable selection method for logistic regression model and compared with results of other methods such as the forward selection, backward elimination, stepwise selection, and AIC statistic.

• Management & Information Systems Review
• /
• v.30 no.2
• /
• pp.211-235
• /
• 2011

We examined the relationship between innovation capability of R&D and the firm's performance by mainly comparing regional strategy industry with non-regional strategy industry. Also, this analysis involved comparing the relationship by regional strategy industry. For the purpose of this study, we divided innovation capability of R&D into input, process and output. The first of main results in this study was that regional strategic industry was significantly higher than non-regional strategy industry in innovation capability of R&D with the exception of the CEO's mind for technological innovations. However, we found no significant difference in the firm's performance. Second, in the results of comparing innovation capability of R&D and the firm's performance by regional strategy industry, electronic-information equipment industry was significantly superior to other industries. Third, it was found that the relationship between innovation capability of R&D and the firm's performance was different by regional strategy industry. Also, R&D manpower and R&D process were more significant factors affecting the firm's performance rather than R&D input and output.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.12
• /
• pp.653-659
• /
• 2018

The accurate prediction of construction cost is a key factor in a project's success. However, it is hard to predict the construction costs in the planning stages rapidly and precisely when drawings, specifications, construction cost calculation statements are incomplete, among other factors. Accurate construction-cost prediction in the planning stage of a project is also important for project feasibility studies and successful completion. Therefore, various techniques have been applied to accurately predict construction costs at an early stage when project information is limited. There are many factors that affect the construction cost prediction. This paper presents a construction-cost prediction method as multiple regression model with seven construction factors as independent variables. The method was used to predict the construction cost of a local housing union project, and the error rate was 4.87%. It is not possible to compare the cost of the project at the planning stage of the local housing union project, but it has high prediction accuracy compared to the unit price of an existing unit area. It is likely to be applied in construction-cost calculation work and to contribute to the establishment of the budget for the local housing union project.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.36 no.6
• /
• pp.507-514
• /
• 2018

TSI (TPI-Slope Index) which is the combination of TPI (Topographic Position Index) and slope was newly proposed for landslide and applied to a landslide susceptibility model. To do this, we first compared the TPIs with various scale factors and found that TPI350 was the best fit for the study area. TPI350 was combined with slope to create TSI. TSI was evaluated using logistic regression. The evaluation showed that TSI can be used as a landslide factor. Then a logistic regression model was developed to assess the landslide susceptibility by adding other topographic factors, geological factors, and forestial factors. For this, landslide-related factors that can be extracted from DEM (Digital Elevation Model), soil map, and forest type map were collected. We checked these factors and excluded those that were highly correlated with other factors or not significant. After these processes, 8 factors of TSI, elevation, slope length, slope aspect, effective soil depth, tree age, tree density, and tree type were selected to be entered into the regression analysis as independent variables. Three models through three variable selection methods of forward selection, backward elimination, and enter method were built and evaluated. Selected variables in the three models were slightly different, but in common, effective soil depth, tree density, and TSI was most significant.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.2
• /
• pp.59-76
• /
• 2023

Since the COVID-19 era, the rise in apartment prices has been unconventional. In this uncertain real estate market, price prediction research is very important. In this paper, a model is created to predict the actual transaction price of future apartments after building a vast data set of 870,000 from 2015 to 2020 through data collection and crawling on various real estate sites and collecting as many variables as possible. This study first solved the multicollinearity problem by removing and combining variables. After that, a total of five variable selection algorithms were used to extract meaningful independent variables, such as Forward Selection, Backward Elimination, Stepwise Selection, L1 Regulation, and Principal Component Analysis(PCA). In addition, a total of four machine learning and deep learning algorithms were used for deep neural network(DNN), XGBoost, CatBoost, and Linear Regression to learn the model after hyperparameter optimization and compare predictive power between models. In the additional experiment, the experiment was conducted while changing the number of nodes and layers of the DNN to find the most appropriate number of nodes and layers. In conclusion, as a model with the best performance, the actual transaction price of apartments in 2021 was predicted and compared with the actual data in 2021. Through this, I am confident that machine learning and deep learning will help investors make the right decisions when purchasing homes in various economic situations.

• Journal of agricultural medicine and community health
• /
• v.38 no.2
• /
• pp.85-96
• /
• 2013

Objectives: This study measures nursing students' ability to recognize child abuse and identifies the factors related to varying levels of recognition. Methods: In this cross-sectional study, data were collected from 370 third and fourth year nursing students using a self-reported questionnaire during November 15-30, 2011. The measuring tool for child abuse recognition used in the study was developed by Ozasa (2011) and is composed of 44 items including physical, mental, and sexual abuse, and neglect. For data analysis, descriptive statistics, two sample t-tests, and regression analysis were evaluated with the SPSS/PC ver20.0 program. Results: Nursing students were concerned about child abuse(85.4%), but knew little about related laws and regulations(14.3%), and they had almost no formal education or training regarding how to recognize child abuse. They only 1.6% reported child abuse even if they encountered such incidents; however, they correctly recognized even infrequent incidents of child abuse. Recognition of sexual abuse ranked highest, while recognition of neglect ranked lowest. Those with higher levels of concern over child abuse showed higher recognition scores. Regression analysis revealed that physical abuse, mental abuse, and neglect had different related factors, while sexual abuse had none. Conclusion: Education and training on the subject of child abuse is strongly recommended in nursing curriculums so that nurses will be able to appropriately respond to and report suspected child abuse.