摘要
提出双方差回归模型,建立双方差回归分析方法,给出其回归方程和高置信水平、高可靠度的置信限曲线。双方差回归模型包含完全相关随机变量和相互独立随机变量,前者可用相关方差表征,后者则需用独立方差描述。传统回归分析主要适用于处理在一条曲线上随机波动的数据,而文中方法则可处理在多条不同曲线上随机波动的数据。在性能曲线测试中,文中方法与成组试验法相比,具有信息量大、精度高,所需试样少的特点。
The two-variance regression model is put forward,on the basis of which the two-variance regression analysis method is established.Then the regression equation and confidence limit curves with high confidence level and high reliability are also given.The two-variance regression model involves totally correlated random variable and independent random variable which can be represented by correlated variance and independent variance,respectively.The presented method extends the regression analysis,which is mainly suitable to test data fluctuating around only one curve,to the test data fluctuating around several different curves that is very common in engineering.Compared with the group test method,the presented method not only has higher precision but also solves the problem of reliability assessment with very small sample.
出处
《机械强度》
CAS
CSCD
北大核心
2011年第1期77-81,共5页
Journal of Mechanical Strength
基金
国家自然科学基金项目(10472006)
北京市教育委员会共建项目(XK100060418)资助~~
关键词
双方差
回归模型
回归分析
可靠性
性能退化
性能测试
Two-variance
Regression model
Regression analysis
Reliability
Performance degradation
Performance test
