• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.788-791
• /
• 2024

온라인 교육의 보급이 증가함에 따라 온라인 학습자의 중도이탈율을 줄이기 위한 효과적인 전략이 필요해 졌다. 본 논문은 인공지능 기술을 온라인 학습 분야에 적용하여 온라인 학습 중도 이탈 문제를 해결하고자 한다. 특히 온라인 교육기관에서 중도이탈예측 시스템을 이미 운영 중인 온라인교육시스템(LMS)을 새롭게 구축하거나 크게 변경하지 않고도 개발 또는 도입이 가능한 방안을 제시한다. 운영 중인 온라인교육시스템(LMS)에 저장되고 있는 온라인 학습자들의 학습자정보 및 학습 이력 데이터를 활용하여 중도 이탈 위험이 있는 학생을 식별하고 맞춤형 관리가 가능하도록 온라인 교육 중도 이탈예측 시스템 개발하고 LMS와 연동하여 서비스는 방법을 제안한다. 이를 통해 온라인교육기관에서는 중도이탈이 있는 학생을 효과적으로 관리함으로써 온라인 교육의 질을 향상시킬 수 있고, 더 많은 수의 온라인 학습자가 성공적으로 과정을 이수할 수 있도록 하는 데 기여할 것이다.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.06a
• /
• pp.421-444
• /
• 2001

최근 데이터 마이닝 기법이 주목받고 있는 이유 중의 가장 큰 이유는 자사가 보유하고 있는 고객의 특성을 파악함으로써 기존의 고객을 효과적으로 유지·관리할 수 있도록 지원하기 때문이다. 특히 고객 보유율 5% 신장이 수익률 120% 증대를 가져오는 것으로 보고되고 있는 신용카드 업계에서는 신규고객을 확보하는 것 만큼 기존 고객을 유지·관리하는 것이 중요하다. 특히, 신용카드를 발급 받고 거의 사용하지 않은 고객이나 쉽게 이탈하는 고객을 판별하는 것은 신용카드사의 입장에서는 비용절감 차원에서 매우 중요하다. 그러나 아직까지 어떠한 속성을 보유하고 있는 고객이 쉽게 이탈하는지를 판별할 수 있는 연구는 거의 진행되지 않았다. 이에 본 인구에서는 데이터마이닝 기법 중 널리 알려진 인공신경망, 로지스틱 회귀분석, C5.0 방법을 이용하여 신용카드 시장에서의 고객현황에 대하여 분석하고자 한다. 이를 위하여 본 연구에서는 모 신용카드사의 최근 4년간 (97넌 3월 이후) 가입고객 및 이탈고객을 대상으로 실증분석을 실시하였다. 분석결과 신용카드 시장에서 카드를 지속적으로 보유하고 있는 고객과 이탈하는 고객을 구분하는 속성이 존재함을 발견하였고, 이를 바탕으로 신용카드사가 수립해야 할 마케팅 전략을 제시하였다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.11
• /
• pp.3465-3472
• /
• 2009

The aims of this study is to medical record data warehouse which had been collected from hospital information systems. continuous patient 2,118 60.5%, informally patient 1,385 39.5%. In using survival factors sex, age, area, insurance, admission-course, medical treatment, out-patient lesson, out-patient form, conference diagnosis, operation, cancer, medical reservation. As a result of making a predictive modeling using the logistic regression, the fitness of the predictive modeling of informally patient was 66.0% and neural network, the predictive was 66.72% and CHAID, the predictive was 63.25%, which is a data mining. The expected modeling of the informally patients, the hospital through the continuous patient management and trust of hospital.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2002.02a
• /
• pp.329-334
• /
• 2002

홍고추 꼭지 제거 장치를 설계·제작하고 실험을 통하여 성능을 검증하였다. 꼭지 제거 시 발생되는 최대 손실은 공급량 48g/s에서 15.5%를 나타냈으며 꼭지 중량 비를 감안한다면 실제 최대 손실율은 약 8.7%로 예상된다. 고추 이탈율은 공급량 87g/s에서 최대 35%를 나타냈으며, 꼭지 제거율은 공급량 48g/s에서 최고 66%를 나타냈다. 총 고추 투입량을 기준으로 환산한 꼭지 제거 수율은 최고 55%로 나타나 당초 예상했던 성능과는 차이를 나타냈다. 공급량 48g/s, 꼭지 제거율 66%, 손실을 28%, 1일 작업 시간 8시간을 기준으로 홍고추 꼭지 제거 능력을 계산하면 약 912kg으로 이는 인력에 의한 처리량의 약 1.5배에 해당한다.

• Land and Housing Review
• /
• v.13 no.2
• /
• pp.91-115
• /
• 2022

This study examines changes in business survival rates in a gentrified commercial district by comparing early movers with late entrants. Using the Hongik University Commerical District, or Hongdae, as a case study, we adopt discrete-time survival analysis to compare survival rates between businesses established before 2000 (early movers) and ones established after 2000 (late arrivers). We compare the business survival patterns in a gentrified commercial district (experimental group) to non-gentrified commercial districts (control group) in Mapogu. We examine a survival-rate difference between early movers and late arrivers by different industrial categories. We finally examine a survival-rate gap between franchise and non-franchised businesses. The results show that the early movers have lower survival rates than the late arrivers in the gentrified Hongdae area, whereas there is no significant difference in survival rates between the early movers and the late arrivers in Mapogu. The early movers in daily-life-supporting businesses in Mapogu have even higher survival rates than the late-arrivers. In addition, franchised businesses have higher survival rates than non-franchised stores both in Hongdae and Mapogu. The results provide statistical and comprehensive evidence of the displacement of early movers at a more rapid pace in gentrified areas than non-gentrified aveas, which has been an anecdotal narrative.

• Communications for Statistical Applications and Methods
• /
• v.18 no.6
• /
• pp.687-696
• /
• 2011

Customer Lifetime or Customer Lifetime Value is a essential metric of differentiated CRM marketing and differentiated marketing strategy as a company core competency. However, customer lifetime used in companies is easily obtained from a confined simple customer attrition rate at some specific time point regardless of customer characteristics. In this study, in order to overcome the constraints of previous simple methods and to make practical use of it in industries, we suggest a method that estimates a customer lifetime using a customer segment based survival analysis with the censored data of customers; in addition, we apply this method to A mobile telecom company data. A method using customer segment based survival analysis is suggested in this study 1) includes all customers having different subscription dates, 2) reduces individual error, 3) can reflect trends after the observed time point and is more realistic.

• Tuberculosis and Respiratory Diseases
• /
• v.42 no.3
• /
• pp.361-369
• /
• 1995

Background: A number of different weaning techniques can be employed such as spontaneous breathing trial, Intermittent mandatory ventilation(IMV) or Pressure support ventilation(PSV). However, the conclusive data indicating the superiority of one technique over another have not been published. Usually, a conventional spontaneous breathing trial is undertaken by supplying humidified $O_2$ through T-shaped adaptor connected to endotracheal tube or tracheostomy tube. In Korea, T-tube trial is not popular because the high-flow oxygen system is not always available. Also, the timing of extubation is not conclusive and depends on clinical experiences. It is known that to withdraw the endotracheal tube after weaning is far better than to go through any period. The tube produces varying degrees of resistance depending on its internal diameter and the flow rates encountered. The purpose of present study is to evaluate the effectiveness of weaning and extubation following a 60 minutes spontaneous breathing trial with simple oxygen supply through the endotracheal tube. Methods: We analyzed the result of weaning and extubation following a 60 minutes spontaneous breathing trial with simple oxygen supply through the endotracheal tube in 18 subjects from June, 1993 to June, 1994. They consisted of 9 males and 9 females. The duration of mechanical ventilation was from 38 hours to 341 hours(mean: $105.9{\pm}83.4$ hours). In all cases, the cause of ventilator dependency should be identified and precipitating factors should be corrected. The weaning trial was done when the patient became alert and arterial $O_2$ tension was adequate($PaO_2$ > 55mmHg) with an inspired oxygen fraction of 40%. We conducted a careful physical examination when the patient was breathing spontaneously through the endotracheal tube. Failure of weaning trial was signaled by cyanosis, sweating, paradoxical respiration, intercostal recession. Weaning failure was defined as the need for mechanical ventilation within 48 hours. Results: In 19 weaning trials of 18 patients, successful weaning and extubation was possible in 16/19(84.2 %). During the trial of spontaneous breathing for 60 minutes through the endotracheal tube, the patients who could wean developed slight increase in respiratory rates but significant changes of arterial blood gas values were not noted. But, the patients who failed weaning trial showed the marked increase in respiratory rates without significant changes of arterial blood gas values. Conclusion: The result of present study indicates that weaning from mechanical ventilation following a 60 minutes spontaneous breathing with $O_2$ supply through the endotracheal tube is a simple and effective method. Extubation can be done at the same time of successful weaning except for endobronchial toilet or airway protection.

• The Korean Journal of Applied Statistics
• /
• v.37 no.4
• /
• pp.525-539
• /
• 2024

Massive multiplayer online role playing game (MMORPG) is a common type of game these days. Predicting user churn in MMORPG is a crucial task. The retention rate of users is deeply associated with the lifespan and revenue of the service. If the churn of a specific user can be predicted in advance, targeted promotions can be used to encourage their stay. Therefore, not only the accuracy of churn prediction but also the speed at which signs of churn can be detected is important. In this paper, we propose methods to identify early signs of churn by utilizing the daily predicted user retention probabilities. We train various deep learning and machine learning models using log data and estimate user retention probabilities. By analyzing the change patterns in these probabilities, we provide empirical rules for early identification of users at high risk of churn. Performance evaluations confirm that our methodology is more effective at detecting high risk users than existing methods based on login days. Finally, we suggest novel methods for customized marketing strategies. For this purpose, we provide guidelines of the percentage of accessed users who are at risk of churn.

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 2004.11a
• /
• pp.306-306
• /
• 2004

• The Korean Journal of Food And Nutrition
• /
• v.11 no.2
• /
• pp.137-142
• /
• 1998

This study aims to examine appropriate kinds of pan oil to release a cake from the pan coated with polyether sulphone and to prove the change of acid value and peroxide value during the process of baking. The major factor effecting the rate of release of the cakes from the pan was properties of the oils rather than the additives. The releasing rate appeared higher as the iodine value was lower. The highest releasing rate was found in the sample, which was made of coconut oil plus 2% of wax, 2% of lecithin. The changes of acid value of the pan oil between the beginning and after baking was little different. In contrast, Peroxide value showed a great change after baking.