• 한국산업정보학회논문지
• /
• 제27권2호
• /
• pp.11-24
• /
• 2022

LMS 기반의 온라인 평가를 위해 출제되는 문제들은 교수자가 직접 출제하거나 또는 카테고리별로 나뉘어진 문제은행에서 난이도에 따른 자동 출제 방식을 주로 이용한다. 이중에서 난이도에 따른 자동출제 방식은 평가자들에게 출제되는 문제가 서로 다를수 있기 때문에 무엇보다 객관적이고 효율적인 방법으로 문제의 난이도를 관리하는 것이 중요하다. 본 논문에서는 문제의 정답률뿐만 아니라 해당 문제를 해결하는데 사용된 소요시간을 같이 고려한 난이도 재조정 알고리즘을 제시한다. 이를 위해 머신러닝의 로지스틱 회귀 분류 알고리즘을 이용하였으며, 학습모델의 예측 확률값을 기반으로 기준 임계값을 설정하여 각 문항별 난이도 재조정에 활용하였다. 그 결과 정답률에만 의존한 문항별 난이도에 많은 변화가 일어남을 확인할 수 있었다. 또한 조정된 난이도의 문제를 이용하여 그룹별 평가를 수행한 결과, 정답률 기반의 난이도 문제에 비해서 대부분의 그룹에서 평균 점수가 향상됨을 확인할 수 있었다.

• 한국인구학
• /
• 제11권1호
• /
• pp.1-13
• /
• 1988

여기서는 많은 수의 비관측사례로부터 발생할 수 있는 표본의 편의(bias) 문제를 탐구한다. 이 연구는 본래 일본 후생성이 1989년 실시한 <가족주기와 가구형태에 대한 인구학적 조사> 자료를 이용하여 노인부보와 자녀간 근접성을 분석하는 목적에서 이루어졌다. 그런데 <가족주기와 가구형태에 대한 인구학적 조사>는 노인부모를 대상으로 한 조사가 아니라 전체 가구 일반에 대한 조사이기 때문에 노인부모에 대한 많은 정보를 손상하고 있었다. 또한 본 조사는 가구주를 통하여 가족원에 대한 정보를 획득하는 방식으로 설계되었기 때문에 가족원에 대한 정보가 완전하지 못하였다. 나아가 비관측사례의 유형을 보면 여러 항목들이 동시적으로 관측되지 않고 있었다. 이와 같이 복합적 메커니즘에서 발생한 비관측 사례는 분석의 편의를 초래할 위험이 크다. 우선, 많은 수의 비관측사례로 표준오차를 잘못 추정할 소지가 크다. 더욱이 사례들이 선택적으로 관측되지 않았다면 관측된 자료에 따른 추정을 심각한 편의를 포함할 수 있다. 이와 같이 손상된 자료로부터 발생할 수 있는 추정 편의를 개선하기 위하여 여기서는 두 가지 기법을 활용하였다. 첫째, 관측치와 공변인간의 관계에 기초하여 비관측사례를 추정하는 방법으로 EM 알고리듬을 활용하였다. 둘째, 관찰의 선택성에서 비롯된 추정 편의를 개선하기 위하여 이단계(two stage) 모형을 활용하였다.

• Applied Biological Chemistry
• /
• 제50권4호
• /
• pp.245-252
• /
• 2007

FTA(fault tree analysis)는 system 오류 관리를 위한 정성적/정량적 기법으로 적용되고 있다. FTA를 적용한 PCR의 오류 관리 system의 구축을 위한 시범적 단계로서 PCR 실행의 여러 단계 중 가장 간단한 단계인 '반응액의 제조 및 PCR 기기 사용 단계'를 모델로 하여 분석하였다. PCR 실행시 발생할 수 있는 오류를 연역적 논리 방식에 의해 fault tree의 형태로 규명하였다. Fault tree는 오류 관리의 최상위 요소인 top event를 중심으로 중간 계층을 이루는 intermediate events와 최하위의 요소인 basic events로 세분하여 구성하였다. Top event는 '반응액의 제조 및 PCR 기기 사용 단계에서의 오류'; 중간계층 events는 '기기 유래 오류', '실험행위 유래 오류'; basic events는 '정전상황', 'PCR 기기 선정', '기기 사용 관리', '기기 내구성', '조작의 오류', '시료 구분의 오류'로 분석되었다. 이로부터 top event의 원인 분석 및 중요 관리점을 도출하기 위하여 정성적/정량적 분석을 실시하였다. 정성적 기법으로 minimal cut sets, structural importance, common cause vulnerability를 분석하였고, 정량적 기법으로 simulation, cut set importance, item importance, sensitivity를 분석하였다. 정성적 분석과 정량적 분석의 결과에서 '시료 구분의 오류'와 '기기 조작의 오류'가 제 1중요관리점; '기기 관리의 오류'와 '내구성에 의한 오류'는 제 2중요관리점으로 일치되게 나타났다. 그러나 '정전상황'과 '기기 선정의 오류'는 정성적 분석에서만 중요관리점으로 분석되었다. 특히 sensitivity 분석에서 '기기 관리의 오류'는 사용 시간이 경과함에 따라 가장 중요한 관리점으로 부각되었다. 결론적으로 FTA는 PCR 모델 case에 대한 오류의 원인 분석 및 그 방지를 위한 중요관리점을 제시함에 따라, 궁극적으로 미래에 PCR의 오류 관리 system을 완성할 수 있는 효과적인 방법으로 사료된다.

• Research in Vestibular Science
• /
• 제17권4호
• /
• pp.152-159
• /
• 2018

Objectives: The Korean Dizziness Handicap Inventory (KDHI), which includes 25 patient-reported items, has been used to assess self-reported dizziness in Korean patients with Parkinson disease (PD). Nevertheless, few studies have examined the KDHI based on item-response theory within this population. The aim of our study was to address the feasibility and clinimetric properties of the KDHI instrument using polytomous Rasch measurement analysis. Methods: The unidimensionality, scale targeting, separation reliability, item difficulty (severity), and response category utility of the KDHI were statistically assessed based on the Andrich rating scale model. The utilities of the orderedresponse categories of the 3-point Likert scale were analyzed with reference to the probability curves of the response categories. The separation reliability of the KDHI was assessed based on person separation reliability (PSR), which is used to measure the capacity to discriminate among groups of patients with different levels of balance deficits. Results: Principal component analyses of residuals revealed that the KDHI had unidimensionality. The KHDI had satisfactory PSR and there were no disordered thresholds in the 3-point rating scale. However, the KDHI showed several issues for inappropriate scale targeting and misfit items (items 1 and 2) for Rasch model. Conclusions: The KDHI provide unidimensional measures of imbalance symptoms in patients with PD with adequate separation reliability. There was no statistical evidence of disorder in polytomous rating scales. The Rasch analysis results suggest that the KDHI is a reliable scale for measuring the imbalance symptoms in PD patients, and identified parts for possible amendments in order to further improve the linear metric scale.

• 복식
• /
• 제63권3호
• /
• pp.17-32
• /
• 2013

This study examined the visual elements of the Trompe l'oeil design and the combination conditions of each element in order to deduce the form generation diagrams of Trompe l'oeil fashion design. The following is what the study found: First, Trompe l'oeil fashion design can be divided into two based on its subject: 'the combination of internal elements of dress' and 'the combination of external elements of dress'. Looking at the visual components of each type, depending on the subject, 'the combination of internal elements of dress' design was achieved by combining internal dress elements: including textile, detail, item, accessory, layering, etc.. In case of 'the combination of external elements of dress' design, the object of Trompe l'oeil was expanded in scope to environment, which includes the actions of a wearer. Second, as for Trompe l'oeil fashion design through the combination of external dress elements, the condition to combined visual components of the work in the process of creation was related to 'the popularity of object', 'the similarity of shapes and sizes', and 'the probability of location'. Third, deriving a form generation diagrams of Trompe l'oeil fashion design was very effective in delineating specific relationships and conditions, directions among elements combined in fashion design.

• 품질경영학회지
• /
• 제24권4호
• /
• pp.207-223
• /
• 1996

These days, customers have attached great importance to the function of product liability and quality assurance. In Korea, the single rectifying sampling inspection for attribute (KS A 3105) has been used. But this inspection plan given by tables (KS A 3105) has some defects. There are limitations in the range of applications and irrationality of approximate probability and the double rectifying sampling inspection is not mentioned. Moreover, ATI (average total inspection) does not reflect sampling costs and the loss of nonconforming item. Therefore, the objectives of this study is to develope new algorithms and computer program that provide the optimal sampling inspection plan based on minimum linear costs (single & double inspection plan). The result of this study revealed that the new algorithm is less than KS A 3105 in ATI and basically, double inspection plan is more economical. Also it comes over restrictions in KS A 3105. So, it is definite that the optimal solution can be obtained considering cost factors in manufacturing and sampling process, and costs can be saved in the long term.

• 경영과학
• /
• 제31권2호
• /
• pp.1-13
• /
• 2014

In almost all of the organizations, the cost for acquiring and maintaining the inventory takes a considerable portion of the management budget, and thus a certain constraint is set upon the budget itself. The previous studies on inventory control for each item that aimed to improve the fill rate, backorder, and the expenditure on inventory are fitting for the commercially-operated SCM, but show some discrepancies when they are applied to the spare parts for repairing disabled systems. Therefore, many studies on systematic approach concept considering spare parts of various kinds simultaneously have been conducted to achieve effective performance for the inventory control at a lower cost, and primarily, METRIC series models can be named. However, the past studies were limited when dealing with the probability distributions for representing the situation on demand and transportation of the parts, with the (S-1, S) inventory control policy, and so on. To address these shortcomings, the Continuous Time Markov Chain (CTMC) model, which considers the phase-type distributions and the (s, Q) inventory control policies to best describe the real-world situations inclusively, is presented in this study. Additionally, by considering the cost versus the system availability, the optimization of the inventory level, based on this model, is also covered.

• Journal of information and communication convergence engineering
• /
• 제15권1호
• /
• pp.43-48
• /
• 2017

Movie ratings are crucial for recommendation engines that track the behavior of all users and utilize the information to suggest items the users might like. It is intuitively appealing that information about the viewing preferences in terms of movie genres is sufficient for predicting a genre of an unlabeled movie. In order to predict movie genres, we treat ratings as a feature vector, apply a Bernoulli event model to estimate the likelihood of a movie being assigned a certain genre, and evaluate the posterior probability of the genre of a given movie by using the Bayes rule. The goal of the proposed technique is to efficiently use movie ratings for the task of predicting movie genres. In our approach, we attempted to answer the question: "Given the set of users who watched a movie, is it possible to predict the genre of a movie on the basis of its ratings?" The simulation results with MovieLens 1M data demonstrated the efficiency and accuracy of the proposed technique, achieving an 83.8% prediction rate for exact prediction and 84.8% when including correlated genres.

• Corrosion Science and Technology
• /
• 제8권3호
• /
• pp.119-125
• /
• 2009

Corrosion assessment has a number of uses but the use considered here is as a precursor to Risk-Based Inspection (RBI) planning. Systematic methods consisting of technical modules of RBI program were used to assess the effect of specific corrosion mechanism on the probability of failure in equipments of petrochemical plants. Especially in part of the damage and corrosion assessment, screening step involved evaluating the combinations of process conditions and construction materials for each equipment item in order to determine which damage mechanisms are potentially active. For general internal corrosion, either API 510 or API 570 was applied as the damage rate in the calculation to determine the remaining life and inspection frequency. In some cases, a measured rate of corrosion may not be available. The technical modules of RBI program employ default values for corrosion, typically derived from published data or from experience with similar processes, for use until inspection results are available. This paper describes the case study of corrosion and damage assessment by using RBI methodology in petrochemical plant. Specifically, this paper reports the methodology and the results of its application to the petrochemical units using the $KGS-RBI^{TM}$ program, developed by the Korea Gas Safety Corporation to suit Korean situation in conformity with API 581 Codes.

• 한국환경과학회지
• /
• 제15권9호
• /
• pp.887-895
• /
• 2006

Through this research of the analysis on the frequency flood discharges regarding basin property factors, a linear regression system was introduced, and as a result, the item with the highest correlation with the frequency flood discharges from Nakdong river basin is the basin area, and the second highest is the average width of basin and the river length. The following results were obtained after looking at the multi correlation between the flood discharge and the collected basin property factors using the data from the established river maintenance master plan of the one hundred twenty-five rivers in the Nakdong river basin. The result of analysis on multivariate correlation between the flood discharges and the most basic data in determining the flood discharges as basin area, river length, basin slope, river slope, average width of basin, shape factor and probability precipitation showed more than 0.9 of correlation in terms of the multi correlation coefficient and more than 0.85 for the determination coefficient. The model which induced a regression system through multi correlation analysis using basin property factors is concluded to be a good reference in estimating the design flood discharge of unmeasured basin.