• Journal of the Korean Data and Information Science Society
• /
• 제27권5호
• /
• pp.1399-1412
• /
• 2016

We compare many normality tests consisting of different sources of information extracted from the given data: Anderson-Darling test, Kolmogorov-Smirnov test, Cramervon Mises test, Shapiro-Wilk test, Shaprio-Francia test, Lilliefors, Jarque-Bera test, D'Agostino' D, Doornik-Hansen test, Energy test and Martinzez-Iglewicz test. For the purpose of comparison, those tests are applied to the various types of data generated from skewed distribution, unsymmetric distribution, and distribution with different length of support. We then summarize comparison results in terms of two things: type I error control and power. The selection of the best test depends on the shape of the distribution of the data, implying that there is no test which is the most powerful for all distributions.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2003년도 추계학술대회 논문집(III)
• /
• pp.363-368
• /
• 2003

The Inter-laboratory comparison test which is demanded on the authorized test of ISO 17025 is the program to guarantee the reliability for test results of test organization. In order to apply to the field of railway, brake disk lining as one of the railway components has been selected and the comparison test is executed for the purpose of guaranteeing the reciprocal reliability for the test result of both organizations and enlarging the substantial technique exchange of both nations by the standard establishment with test procedures and test programs from now on.

• The Journal of Korean Physical Therapy
• /
• 제15권4호
• /
• pp.479-487
• /
• 2003

Tje purpose of this study is to understand body as a whole concept through examining effects of distant shoulder and neck with GCM ankle joint therapy. The twenty three subject with ankle joint therapy and comparison groups without ankle joint therapy were authorized the score of before and after therapy by correspondence T-test using pump lines and tapes as instruments. The results are as follows. 1. The test group had change of shoulder in measurement 1, 2, 3 and were regarded statistically 1st measurement on left, 2nd on right.(p<.05) 2. The test group had change of neck in measurement 1, 2, 3 and were regarded statistically 2nd on left on right.(p<.05) 3. In the comparison group the comparison of shoulder height was not regarded statistically all at 1, 2 comparison on left and regarded statistically all at 1, 2 comparison on right.(p<.05) 4. In the comparison group the comparison of neck height was not regarded statistically both left and right at 1, 2 comparison.(p<.05)

• 펄프종이기술
• /
• 제48권1호
• /
• pp.19-26
• /
• 2016

In the consumer products industry, it has been highly desirable to develop objective test methods to replace subjective evaluation methods. In developing an objective test method, subjective evaluation data should be on a linear scale. According to Thurstone's theory of comparative judgment, a%-preference from a paired-comparison test can be converted to a linear-scale value. The required number (N) of paired-comparison tests increases dramatically as the number of products increases. This problem should be solved by classifying the total products into several subgroups consisting of 3-4 products in each group. By doing so, it can not only significantly reduce the number of required paired-comparison tests, but it can also obtain more reliable, reproducible data.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 제36회 하계학술대회 논문집 C
• /
• pp.2217-2219
• /
• 2005

In this paper, an inter-comparison test of lightning impulse reference measurement system between Korea and Japan is reported. At present, such a direct comparison test as the activity to sustain the traceability of the measurement system is generally proposed. Due to the inter-comparison test, ascertaining the grade of the measurement uncertainty and maintaining the level of reference system are accomplished.

• Journal of applied mathematics & informatics
• /
• 제5권1호
• /
• pp.263-268
• /
• 1998

This paper proposes the two-sample comparison us-ing sign test based on ranked-set sample(RSS). We investigate the asymptotic properties of the proposed test statistic and compare the asymptotic relative efficiencies of the proposed test statistic with re-spect to Mann-Whitney-Wilcoxon test statistic based on RSS and Mann-Whitney-Wilcoxon test statistic based on the simple random sample(SRS).

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2012년도 춘계학술대회 논문집
• /
• pp.367-368
• /
• 2012

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2004년도 추계학술대회 논문집
• /
• pp.483-488
• /
• 2004

The inter-laboratory comparison test which is demanded on the authorized test of ISO/IEC 17025 is the program to guarantee the confidence of test result. This paper, as a part of the inter-laboratory comparison test between South Korea and China, is refered to the characteristic of friction coefficient of disk brake lining for rolling stock according to disk type. Brake tests were carried out under constant brake force and operating sequence by using dynamo-tester according to disk and disk lining types. To establishing the confidence of test result, we calculated A type uncertainty about friction coefficient and investigated the factors about the variation of friction coefficient.

• 응용통계연구
• /
• 제19권2호
• /
• pp.231-240
• /
• 2006

일반적으로 쌍체 비교(paired comparison 또는 paired t test)가 효과적인 쌍으로의 구획을 통하여 소위 독립비교(group comparison)라고 일컫는 독립 이표본 t 검정(two-sample t test)보다 정밀한 결과를 제공하는 것으로 알려져 있다. 본 논문에서는 이와 같은 사실을 통계적으로 됫받침하기 위하여 두 방법의 정밀도를 보다 정량적으로 비교하였다. 독립 비교와 쌍체 비교가 각각 일원배치법과 이원배치법 (특히, 임의화 블록 설계법)의 특수한 경우이므로 각각에 상응하는 신뢰구간의 길이의 비로 정의된 비교 통계량은 분산분석의 통계량으로 표현된다. 다시 말해서, 이 통계량은 t 와 F 분포함수로 표현되므로 쉽게 독립비교와 쌍체비교의 정밀도를 비교할 수 있다. 실제로, 자주 쓰이는 유의수준에서는 쌍체표본 내에 구획 간의 유의한 변동이 있으면 쌍체비교가 독립비교보다 정밀할 가능성이 더 높다.

• 한국태양에너지학회 논문집
• /
• 제29권2호
• /
• pp.70-75
• /
• 2009

In this experiment, we did sampling 6 kinds of photovoltaic modules and analyzed the discrepancy of measurement results between l laboratory and 4 PV makers to have performance repeatability at Standard Test Condition(STC) condition. From the KIER's results, Korea's standard test laboratory, other laboratory showed -10% measurement variation. The causes came from correction of reference cell, test condition and the state of skill. Form the comparison test, we analyzed the problems. But three PV maker reduced measurement variation, other one PV maker and one test laboratory didn't improve the problems of correction of reference cell, test condition and the state of skill. Also, High Efficiency Module had a big discrepancy of -10.0$\sim$-6.2% among 3 laboratories which have a less than 10msec light pulse duration time. This made low spectrum response speed so the Fill Factor decreased maximum output power under 10msec light pulse duration time