| Project/Area Number |
07680475
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
社会システム工学
|
|---|
| Research Institution | MUSASHI INSTITUTE OF TECHNOLOGY |
|---|
Principal Investigator |
YOKOYAMA Shinichiro MUSASHI INSTITUTE OF TECHNOLOGY,ENGINEERING,PROFESSOR, 工学部, 教授 (50158375)
|
|---|
| Project Period (FY) |
1995 – 1996
|
| Project Status |
Completed (Fiscal Year 1996)
|
| Budget Amount *help |
¥1,500,000 (Direct Cost: ¥1,500,000)
Fiscal Year 1996: ¥700,000 (Direct Cost: ¥700,000)
Fiscal Year 1995: ¥800,000 (Direct Cost: ¥800,000)
|
|---|
| Keywords | STATISTICAL DATA ANALYSIS / SMALL SIZE OF DATA / OUTLIER / WELBULL DISTRIBUTION / PARAMETER ESTIMATION / BOOTSTRAP METHOD / PRACTICAL USE / 加速寿命試験法 / 異常値(外れ値) / モーメント法 / ワイブル分布 |
|---|
| Research Abstract |
In recent year, an environment problems and problems of the safety of product come to appear. A social responsibility to the enterprise and the activity of consumer protection are increased. Under such a background, it is necessary that reliability and safety of products are analyzed with small size of data. On the one hand, the Weibull distribution is widely used in life testing and reliability theory. In the reliability analysis of filed data, it is well known that there are two difficult situations when we try to estimate Weibull parameters. It is one case of difficulties that number of sample is very small. The other one is presence of outliers. Our tools to solve these problems are called "Bootstrap Parameter Estimation Method" and "DS Outliers Detection Method".
Bootstrap Parameter Estimation Method is proposed by authors, and this method creates the robust estimator for outliers and also this estimator provide high accuracy when the sample size is very small. The practical outliers discordancy test for Weibull population is not proposed. Then we propose a new outliers discordancy procedure i.e., DS outliers detection method. The detection step of this procedure is very simple, and this procedure shows the high power even for the case when many outliers are present. In this research it is described that removing the outliers gives more accurate estimates than using the robust estimator.
Furthermore, in this research the new procedure of Weibull parameter estimation with suspended data is proposed. This procedure has 2 steps to get the accurate estimator. It is described that proposed 2-step estimation method gives good accuracy than the usual estimators.
|
