Earticle

A fast statistical method for Chinese unknown word detection is proposed. It is based on association measure, heuristic features and LocalMaxs algorithm. The unknown words discovered are not limited to certain word patterns. It is also effective for low frequency unknown words and sensitive to new coming texts. Experiment shows its effectiveness.

Since the democratization of powerful AI engines for the game of Go, it is not uncommon to see a drastic level increase of some players that must be explained with the help of AI. This is considered cheating and forbidden by most organizations. When looking at online beginners and stronger amateur players, we discovered that they can display playing strength below professional level and still confidently win the game, as opposed to professional players. This makes using only AI-likeness metrics not sufficient to detect such players. We propose a method based on the analysis of a player’s performance considering point loss distribution over several games, taking into account only relevant moves of a game. We still use an AI-likeness metric for analyzing individual games where the use of AI may not be consistent. We evaluated our methods on two European go official online leagues, where cheating detection was already performed (for a total of about 150 unique regular players, with levels ranging from 20 kyu to 5 European dan). We show that our system confirmed 5 cases of players previously banned for cheating (out of 6). Our methods do not set out to categorize players between “cheaters” and “not cheaters,” but rather rank them in order of suspicion, for the sake of assisting referees and providing them a way to effectively investigate suspicious players over time.

본 연구의 목적은 다양한 가중치 산정방법을 이론적 그리고 실증적으로 비교 및 분석함으로써 최적의 세부지표 가중치 산정방법을 제시하는 것이다. 이를 위해 본 연구에서는 가중치 산정방법을 지수산정기관 또는 연구자가 임의로 가중치를 산정하는 주관적 방법, 회귀분석, 요인분석 등의 통계분석모형을 이용하는 통계적 분석방법, 그리고 전문가들의 의사결정에서의 불확실성을 적절하게 고려할 수 있는 사회적 판단방법으로 구분하고, 이들 다양한 방법들의 논리적 타당성, 측정의 용이성, 가중치 산정 결과의 유사성 등을 비교 분석하였다. 본 연구의 분석결과는 지표간의 상대적 중요도를 가중치로 반영하는 방법 중에서 가중치 산정이 객관적이고 체계적이라고 판단되는 계층적 분석과정 또는 퍼지집합이론이 가장 우수한 것으로 나타났다. 그러나 계층적 분석과정의 경우 세부지표의 수가 많아지면 단순화된 계량방법을 사용한다고 하더라도 설문항목이 많아지는 경향이 있다. 이러한 단점을 해결하기 위해 다수의 세부지표를 단순화하는 방법을 병행하여 가중치 산정과정을 진행하는 것이 바람직하다고 판단된다. 다수의 세부지표를 소수의 유사지표군으로 단순화시키는 과정에서는 연구자 또는 전문가의 주관적 판단이나 요인분석을 적용하고, 소수의 유사지표군에 대해서는 계층적 분석과정이나 퍼지집합이론을 적용하여 세부지표의 가중치를 산정할 것을 제안한다.

This study proposes an optimal indicator weight calculation method for science and technology innovation capability index. Three most widely used weight calculation methods such as subjective method, statistical method (regression analysis or factor analysis), and expert judgement method(analytic hierarchy process or fuzzy set theory) are analyzed. Our analyses show that expert judgement method is better than others because its weight calculation process is the most objective and systematic although it becomes very complex and cumbersome to calculate the weights when there are a large number of indicators. Thus, we suggest a two stage weight calculation scheme; first, apply factor analysis to reduce a large number of indicators into several groups of indicators, and thereafter, use analytic hierarchy process or fuzzy set theory to calculate the indicator weights for a small number of groups of indicators.

A quantified reasoning and description of the perplexity for evaluating language models by using the concept of information entropy is discussed in this article: The smaller the entropy of the language estimated by the language model is, the more precise the language model is; an interpolated model based on two (n‐1)‐gram models is better than the (n‐1)‐gram component models, but not a n‐gram model. We also explore the methods to estimating the entropy of Chinese using language models.

Purpose: The purpose of this study was to investigate the relationship and the most influencing sub-variable of emotional labor significantly predicts to burnout and job satisfaction among clinical nurses. Methods: This descriptive study was conducted with a convenient sample 271 of clinical nurses in K providence, Korea. Surveyed data was collected from April 17, 2010 to May 14, 2010. To determine the most effective variable of emotional labor to burnout and job satisfaction, the dada were analyzed by Canonical correlation using SAS 9.1. Results: Canonical correlation analysis revealed that emotional intensity among 5 sub-domains of emotional labor made the most outstanding contribution in predicting to burnout. Emotional deep acting was the most outstanding contribution in predicting to job satisfaction of clinical nurses. Conclusion: The results suggest that reducing emotional intensity should be included in structured burnout management. Especially the positive reinforcement of emotional deep acting for clinical nurses is important for increasing job satisfaction.

Background: The use of the extracts from evening primrose seeds as a health functional food has been gradually increased. Therefore, the monitoring and screening process has been considerably required for its quality control. Objective: This study aimed to estimate the measurement uncertainty associated with determination of penta-o-galloyl β-D-glucose (PGG) in extracts from evening primrose seeds by high-performance liquid chromatography. Methods: The sources of measurement uncertainty was expressed in accordance with mathematical/statistical theories of GUM & EURACHEM Guide. The expanded uncertainty was calculated by using the relative standard uncertainty between analytical result and sources of uncertainty in measurement (sample weight, final volume, extraction volume, standard solution, matrix and instrument etc). Results: In the results of 95% confidence interval, the uncertainty in measurement was 10,253.34 ± 1,844.50 μg/kg (k = 2.0). Conclusion: In this study, it showed that the value of uncertainty in measurement for determination of PGG in extracts from evening primrose seeds by HPLC has about 18.0% influence on PGG contents of the analytical results. The results would be very useful for the monitoring and screening of evening primrose seeds marketed in Korea for its quality control as dietary supplement.

서비스수준관리(Service Level Management)의 목적이 단순한 재무적 계약(서비스 수준협약의 계약 자체)에서 고객의 기대와 제공자의 서비스에 대한 관리 시스템(서비스 수준협약에 대한 관리시스템)으로 최근 변해가고 있다. 본 논문의 목적은 통계적 기법을 이용해서 체계적인 정보기술 서비스 수준관리 방안을 제시하는 것이다. 본 연구는 통계적 기법과 서비스 수준관리 라는 두 가지 관점에서 서비스 수준을 다루고 있다. 첫째, 통계적 기법을 이용해서 목표 서비스 수준을 설정한다. 둘째, 전체 서비스 수준에 대한 평가방법을 제시한다. 이와 같이 제시된 방법을 이용해서 기업 전략 측면에서 정보기술 아웃소싱 성과로서 종합 서비스 수준을 계량적으로 측정하는 것이 가능하다.

The purpose of IT Service Level Management(SLM) has recently shifted from being a simple financial contract(SLA contract itself) towards a system(SLA management system) for the management of the customer's expectations and the provider's service. The aim of this paper is to introduce and present the systematic SLM using statistical method. We discuss the service level according to two points of view: statistical method and service level management. First, we go into more detail on the objective service level using statistical method. Second, we give an overview of total service level assessment method. The proposed framework enables to measure total service level as IT outsourcing performance quantitatively from the aspect of business strategy.

Among many potential problems associated with automobiles, the problem regarded as most sensitive and important to the customers is the automobile’s quietness. In particular, because the trend toward the high-performance automobiles is increasing, the quietness should be considered reflecting the increasing demand of the customers for the comfort of automobile ride. Road noise is the low-frequency range in the region of 20Hz to 350Hz, generated when the tires are making contacts with the rough road surface while the automobile is running, in which the noises are resonating within the passenger compartment or with the automobile suspension. In this research, among the factors related to the road noise, the influencing factors associated with the suspension and the tire design were investigated. In particular, among the tire design factors, the structure and the rubber material were investigated as the influencing factors, and for the investigative purpose, experimental tires were manufactured and road noise evaluations were conducted thorough actual feeling tests. Then the test results and correlations were analyzed using the 6-Sigma statistical method. Results from this study shows that, among the tire design factors, the apex is the most influencing factor, in which the height of the apex is the key. The influential factors were found out to be apex height, belt angle, and tread hardness, in that order, and when the suspension is of a same kind, the tendency of the road noise was found to be similar.

공공부문 경영현장에서는 성과관리를 위한 도구로 균형성과표(Balanced Score Card : BSC)를 도입하는 기관이 늘어가고 있다. Kaplan & Norton에 의해 발표된 후 새로운 경영정보시스템이자 전략정보시스템으로 조직의 성과를 향상시키는 잠재력을 갖춘 중요한 경영도구로써 빠르게 확산되어 왔다. 지금까지 대부분의 조직에서는 구성원들의 토론을 거치는 등의 방법으로 BSC 관점과 전략목표를 도출하여 왔다. 그러나 이러한 방법은 과학적이고 합리적이지 못해 성과관리 시스템의 실패요인으로 지적 받아왔다. 본 연구는 요인분석 등의 통계학적인 방법으로 관점을 도출하고 이를 구성원들의 토의과정을 통하여 성과관리에 중요하다고 생각하는 관점을 도출하고 구성원들이 결정하는 브레인스토밍(brainstorming)을 통하여 확정하였다. 그리고 계층적 분석기법(Analytic Hierarchy Process : 이하 AHP)을 적용하여 관점 및 전략목표에 대한 가중치를 산정하고 이를 토대로 구성원들의 토의과정을 거쳐 최종적으로 관점과 가중치를 결정하는 과정을 제시하고 있다. 이를 통하여 구성원이 모두 동의할 수 있는 합리적인 BSC 성과관리 시스템을 구축하였다.

Public organizations have attempted to adopt BSC(Balance Score Card) as a tool for performance management. Since its introduction in 1992 by Kaplan and Norton who recognized the problems and vagueness of existing performance management systems, BSC has spread at a rapid pace as new management information and strategic information systems and as an important management tool with the potential of improving organization performance. They derive BSC perspectives and strategic objectives through discussions of organization members. However, Some research pointed out that this kind of method was a cause of BSC fail. This study is a case study of S public corporation that applies Statistical method, brainstorming and AHP to derive BSC perspectives and strategic objectives. This method creates parametric and reasonable results and makes a successful performance management system.

Objectives: This study examines the spatial and temporal patterns of hypertension prevalence and treatment, while also suggesting a public health intervention strategy for hypertension management. Methods: We utilized the hypertension prevalence and treatment rate from the Community Health Survey data collected between 2011 and 2020. To analyze the spatio-temporal patterns of hypertension prevalence and treatment, spatial autocorrelation and hotspot analyses were performed. Results: The spatial autocorrelation analysis revealed regional hypertension prevalence clusters between adjacent regions. For the previous five years, the hotspots with a high prevalence of hypertension were 28 si-gun-gu, which were adjacent to each other, including 16 regions in Gangwon, four in Gyeonggi, two in Incheon, and six in Chungnam. The treatment rate of hypertension showed a random pattern, irrespective of regional and spatial proximity. Conclusions: The results highlight the cooperative approach from adjacent regions by reflecting regional and spatial effects for effective hypertension management. In addition, it is necessary to implement customized interventions for regions with a low treatment rate.

일반적으로 학생 오케스트라 활동은 창의력과 감수성을 개발시켜 청소년기의 학생들에게 올바른 취미생활을 제공함으로써 사회적・정서적 불안감을 해소시켜 주고 소속감을 갖는 등 안정감을 주는 것으로 알려져 있다. 이에 본 연구에서는 대학생들에게도 오케스트라 동아리 활동이 비슷한 결과를 야기하는지에 대하여 알아보았다. 그 결과 청소년기의 학생들과는 달리, 오케스트라 동아리 활동 자체에 대한 성취도는 높으나, 사회성이나 자신감 향상에는 크게 도움이 되지는 못하는 것으로 나타났다. 그러함에도 불구하고 긍정적인 요소들도 있고 또한 빅데이터를 활용한 인식분석(오피니언 마이닝) 등을 통해 보았을 때 분명히 긍정적인 요소가 있는 만큼, 대학생 오케스트라 활동에 대해서도 국가차원의 지원이 있어야 하겠다.

General student orchestra activity is known as desirable hobby for students of adolescent period, developing their creativity and sensitivity, give students sense of belonging and stability by resolving their social, emotional anxiety. Accordingly, his research investigated whether orchestra club activity also has similar effect on university students. As a result unlike of adolescent students, orchestra activity turned out to be not that helpful for the social, self-confidence improvement of university students, though achievement of the activity itself was high. Despite of the result, there exist positive factors; obviously the activity has positive factors analyzing through recognition analysis (opinion mining) using big data. Therefore national support is required also for the orchestra activity of the undergraduates.

본 연구는 문제은행방식으로 운영되고 있는 국가기술자격 필답형 실기시험문제의 문제추출 을 위한 선정기준에 통계분석법을 활용하여 종목별 특성을 반영하고자 진행하였다. 이를 위해 수험인원 및 시행횟수가 많은 한 종목을 대상으로 선정, 최근 합격률 추세와 수험자 평균점수 데이터를 분석하여 적정 수준의 평균점수 및 합격률을 설정하였다. 또 최근 2년간 시행한 문 항에 대하여 문항별 수험자 평균득점에 영향을 미치는 요인을 다섯 가지 - 문항의 유형, 난이 도, 형태, 재출제, 출제기준항 - 로 지정하고, 요인별 평균득점간 유의성을 분산분석법으로 검 증하였으며, 선정된 주요인의 양적 조정을 통하여 해당 종목에 최적화된 선정기준안을 도출하 였다. 도출된 선정기준안은 2015년도 정기 실기시험에 적용ᆞ교부하여 목표평균점수와 실제 수험자 평균득점간 일치성을 검증하였으며, 본 연구 결과를 바탕으로 필답형 실기시험으로 시 행되는 국가기술자격의 종목별 특성을 반영한 문항관리기법을 제시하였다.

The purpose of this study is development of ‘the criteria for selection’ for make up open-ended type test questions operated by item pool method for national technical qualification in Korea. In order to develop the criteria for selection from item pool, the examinee’s response characteristics of a subject analyzed by using statistical method. This subject was selected by some specific options, which are the number of people and frequencies of test. The item analysis of this subject was carried out using last two years regular test data. Although there are many different factors that affect average score of examinees, in this case, five factors - item types, item difficulties, item categories, the ratio of questions from previous test, question making standard- were chosen for reviewing the interrelationship between average score and each factors. Through the conclusions of analysis of variance, the new criteria draft was reached. The derived new criteria draft in this study was applied to the regular test on 2015. Based on the results, the test item management mechanism considering examinee’s response characteristics was suggested.