• Communications for Statistical Applications and Methods
• /
• 제23권4호
• /
• pp.355-362
• /
• 2016

Ensemble methods often help increase prediction ability in various predictive models by combining multiple weak learners and reducing the variability of the final predictive model. In this work, we demonstrate that ensemble methods also enhance the accuracy of prediction under kernel ridge regression and kernel logistic regression classification. Here we apply bagging and random forests to two kernel-based predictive models; and present the procedure of how bagging and random forests can be embedded in kernel-based predictive models. Our proposals are tested under numerous synthetic and real datasets; subsequently, they are compared with plain kernel-based predictive models and their subsampling approach. Numerical studies demonstrate that ensemble approach outperforms plain kernel-based predictive models.

• 보건교육건강증진학회지
• /
• 제30권3호
• /
• pp.25-34
• /
• 2013

Objectives: The main purpose of this study was to examine factors influencing school bullying victimization of middle school students in relation to social support, self-esteem, stress coping behavior, and health status. Methods: The questionnaire survey was carried out on a convenience sample of 441 middle school students. The data analysis procedure included frequency, ${\chi}^2$-test, t-test, and multiple logistic regression. Results: It was found that 18% of the subjects were bullied by other students. Multiple logistic regression analysis revealed that the factors such as sex(OR=2.35, p=.006), aggressive coping behavior(OR=1.18, p=.028), and health status(OR=1.04, p=.002) were significant affecting factors. Conclusions: The findings suggest that to prevent middle school students' bullying victimization, it is necessary to design intervention programs that considering their health status and stress coping behavior.

• 대한산업공학회지
• /
• 제36권1호
• /
• pp.13-21
• /
• 2010

The prediction of contract cancellation of customers is essential in insurance companies but it is a difficult problem because the customer database is large and the target or cancelled customers are a small proportion of the database. This paper proposes a new data mining approach to the binary classification by handling a large-scale unbalanced data. Over-sampling, clustering, regularized logistic regression and boosting are also incorporated in the proposed approach. The proposed approach was applied to a real data set in the area of insurance and the results were compared with some other classification techniques.

• Asian Pacific Journal of Cancer Prevention
• /
• 제13권1호
• /
• pp.309-314
• /
• 2012

Objective: To investigate the psychological characteristics of hepatic malignancy patients before interventional procedures and assess associations with related factors. Methods: Two hundred and thirteen patients requiring interventional procedure for hepatic malignancy were asked to complete a survey of health knowledge and psychological symptom on health knowledge questionnaire and SCL-90 before interventional procedure. Logistic regression analysis was employed to determine the association of various demographic, clinical and health knowledge factors with the presence of psychological symptoms in patients. Results: Eight psychological symptom scores, i.e. somatization, obsessive-compulsive tendencies, depression, anxiety, hostility, phobia, paranoid ideations and psychotic states, were significantly higher than the normal range (P< 0.001). Of 213 cases in the study, 49 families (23.00%) concealed the diagnoses of hepatic carcinoma from patients; 135 patients (63.38%) described the prognosis of the disease correctly. It was demonstrated that the correlations between psychological symptoms and related factors, i.e. age, gender, education, interventional procedure times and health knowledge, were statistically significant (P<0.05). Conclusion: Psychological distress is severe in hepatic malignancy patients before interventional procedures. Age, gender, education, interventional procedure times and health knowledge are associated with psychological symptoms which are significant different from the normal range in Chinese.

• 응용통계연구
• /
• 제23권5호
• /
• pp.857-870
• /
• 2010

국립축산과학원에서는 2006년 부터 2008년 까지 전국 소비자들을 대상으로 한우 거세우 표본 시료에 대한 관능 평가 조사를 실시하여 데이터를 수집하였으며 본 연구에서는 한우 관능 평가 데이터에 대해 사회 인구학적 요인과 한국 소비자들의 맛 평가에 대한 연관성을 탐구하고자 한다. 소비자 거주지역, 연령, 성별, 직업, 월수입과 쇠고기 부위를 설명변수로 맛등급 평가를 반응변수로 이항 다중 로지스틱 모형과 다항 다중 로지스틱 모형을 적합하고 회귀계수별 유의성 검정과 적합도 검정을 실시한다. 단계별 변수 선택으로 최종 모형을 선택하고 반응변수 범주에 대한 오즈비를 계산하여 맛등급과 설명변수들 간의 관련성을 파악한다. 또한 맛과 관련 있는 연속형 변수를 설명변수로 포함한 경우에 대해서도 이항 다중 로지스틱 모형과 다항 다중 로지스틱 모형을 적합하고 비교한다. 그 결과 거주 지역, 연령, 월수입과 쇠고기 부위 변수들이 선택되었으며 영남지역에서 맛에 대한 오즈가 큰 편이며 수입이 많고 연령이 높을수록 맛에 대한 오즈가 작은 편이었다. 요리법으로는 탕에 대한 구이의 오즈비가 큰 편이며 쇠고기 부위별로는 우둔에 비해서 등심이 다른 부위들 보다 맛에 대한 차이가 크다고 볼 수 있다. 연속형 변수로는 연도가 맛등급에 큰 영향을 미치는 변수로 나타났다.

• Journal of the Korean Data and Information Science Society
• /
• 제27권4호
• /
• pp.911-923
• /
• 2016

빅데이터 시대에 머신러닝의 중요성은 더욱 부각되고 있고 로지스틱 회귀는 머신러닝에서 분류를 위한 방법으로 의료, 경제학, 마케팅 및 사회과학 전반에 걸쳐 널리 사용되고 있다. 지금까지 R과 Hadoop의 통합환경인 RHIPE 플랫폼은 설치 및 MapReduce 구현의 어려움으로 인해 거의 연구가 이루지 지지 않았다. 본 논문에서는 대용량 데이터에 대해 로지스틱 회귀 추정을 위한 두가지 알고리즘 즉, Gradient Descent 알고리즘과 Newton-Raphson 알고리즘에 대해 MapReduce로 구현하고, 실제 데이터와 모의실험 데이터를 가지고 이들 알고리즘 간의 성능을 비교하고자 한다. 알고리즘 성능 실험에서 Gradient Descent 알고리즘은 학습률에 크게 의존하고 또한 데이터에 따라 수렴하지 않는 문제를 갖고 있다. Newton-Raphson 알고리즘은 학습률이 불필요 할 뿐만 아니라 모든 실험 데이터에 대해 좋은 성능을 보였다.

• Communications for Statistical Applications and Methods
• /
• 제22권6호
• /
• pp.655-664
• /
• 2015

A nonresponse adjusted raking ratio estimator that consists of weighting adjustment using estimated response probability and raking procedure is often used to reduce the nonresponse bias and keep the calibration property of the estimator. We investigated asymptotic properties of nonresponse adjusted raking ratio estimator and proposed a variance estimator. A simulation study is used to examine the performance of suggested estimators.

• 품질경영학회지
• /
• 제29권3호
• /
• pp.124-137
• /
• 2001

In general, it is very efficient and effective to use screening variables that are correlated with the performance variable in case that measuring the performance variable is impossible (destructive) or expensive. The general methodology for searching surrogate variables is regression analysis. This paper considers the inspection problem in CRT (Cathode Ray Tube) production line, in which the performance variable (dependent variable) is binary type and screening variables are continuous. The general regression with dummy variable, discriminant analysis and binary logistic regression are considered. The cost model is also formulated to determine economically inspection procedure with screening variables.

• Journal of the Korean Data and Information Science Society
• /
• 제19권2호
• /
• pp.609-620
• /
• 2008

There may be several parallel equipments having the same function in a multi-stage manufacturing process, which affect the product quality differently and have significant differences in defect rate. The product quality may depend on what equipments it has been processed as well as what process variable values it has. Applying one model ignoring the presence of different equipments may distort the prediction of defect rate and the identification of important quality variables affecting the defect rate. We propose a procedure for data segmentation when constructing models for predicting the defect rate or for identifying major process variables influencing product quality. The proposed procedure is based on the principal component analysis and the analysis of variance, which demonstrates a better performance in predicting defect rate through a case study with a PDP manufacturing process.

• 한국신뢰성학회:학술대회논문집
• /
• 한국신뢰성학회 2001년도 정기학술대회
• /
• pp.277-278
• /
• 2001

The successful operation of a product In service depends upon the effective provision of logistic support in order to achieve and maintain the required levels of performance and customer satisfaction. Logistic support encompasses the activities and facilities required to maintain a product (hardware and software) in service. Logistic support covers maintenance, manpower and personnel, training, spares, technical documentation and packaging handling, storage and transportation and support facilities.The cost of logistic support is often a major contributor to the Life Cycle Cost (LCC) of a product and increasingly customers are making purchase decisions based on lifecycle cost rather than initial purchase price alone. Logistic support considerations can therefore have a major impact on product sales by ensuring that the product can be easily maintained at a reasonable cost and that all the necessary facilities have been provided to fully support the product in the field so that it meets the required availability. Quantification of support costs allows the manufacturer to estimate the support cost elements and evaluate possible warranty costs. This reduces risk and allows support costs to be set at competitive rates.Integrated Logistic Support (ILS) is a management method by which all the logistic support services required by a customer can be brought together in a structured way and In harmony with a product. In essence the application of ILS:- causes logistic support considerations to be integrated into product design;- develops logistic support arrangements that are consistently related to the design and to each other;- provides the necessary logistic support at the beginning and during customer use at optimum cost.The method by which ILS achieves much of the above is through the application of Logistic Support Analysis (LSA). This is a series of support analysis tasks that are performed throughout the design process in order to ensure that the product can be supported efficiently In accordance with the requirements of the customer.The successful application of ILS will result in a number of customer and supplier benefits. These should include some or all of the following:- greater product uptime;- fewer product modifications due to supportability deficiencies and hence less supplier rework;- better adherence to production schedules in process plants through reduced maintenance, better support;- lower supplier product costs;- Bower customer support costs;- better visibility of support costs;- reduced product LCC;- a better and more saleable product;- Improved safety;- increased overall customer satisfaction;- increased product purchases;- potential for purchase or upgrade of the product sooner through customer savings on support of current product.ILS should be an integral part of the total management process with an on-going improvement activity using monitoring of achieved performance to tailor existing support and influence future design activities. For many years, ILS was predominantly applied to military procurement, primarily using standards generated by the US Government Department of Defense (DoD). The military standards refer to specialized government infrastructures and are too complex for commercial application. The methods and benefits of ILS, however, have potential for much wider application in commercial and civilian use. The concept of ILS is simple and depends on a structured procedure that assures that logistic aspects are fully considered throughout the design and development phases of a product, in close cooperation with the designers. The ability to effectively support the product is given equal weight to performance and is fully considered in relation to its cost.The application of ILS provides improvements in availability, maintenance support and longterm 3ogistic cost savings. Logistic costs are significant through the life of a system and can often amount to many times the initial purchase cost of the system.This study provides guidance on the minimum activities necessary to Implement effective ILS for a wide range of commercial suppliers. The guide supplements IEC60106-4, Guide on maintainability of equipment Part 4: Section Eight maintenance and maintenance support planning, which emphasizes the maintenance aspects of the support requirements and refers to other existing standards where appropriate. The use of Reliability and Maintainability studies is also mentioned in this study, as R&M is an important interface area to ILS.