青岛理工大学学报
青島理工大學學報
청도리공대학학보
JOURNAL OF QINGDAO TECHNOLOGICAL UNIVERSITY
2012年
1期
126-130
,共5页
线性回归分析%点到直线距离%残差%误差因素
線性迴歸分析%點到直線距離%殘差%誤差因素
선성회귀분석%점도직선거리%잔차%오차인소
linear regression analysis%distance from test point to line%residual%error factors
工程试验中常需进行2个变量间的回归分析，现行最小二乘法通过函数值残差最小而求得拟合曲线，但是误差因素分析表明自变量往往也存在误差，因此提出点到直线距离平方和最小的拟合方法，并给出了数字例子，计算结果与最小二乘法相近．
工程試驗中常需進行2箇變量間的迴歸分析，現行最小二乘法通過函數值殘差最小而求得擬閤麯線，但是誤差因素分析錶明自變量往往也存在誤差，因此提齣點到直線距離平方和最小的擬閤方法，併給齣瞭數字例子，計算結果與最小二乘法相近．
공정시험중상수진행2개변량간적회귀분석，현행최소이승법통과함수치잔차최소이구득의합곡선，단시오차인소분석표명자변량왕왕야존재오차，인차제출점도직선거리평방화최소적의합방법，병급출료수자례자，계산결과여최소이승법상근．
Regression analysis is often needed in engineering test. Fitted curve is got through current least square method as the residual function value is minimum. Error factor analysis shows that the error still exists in independent variables. This paper suggests a new curve fitting method that quadratic sum of the distance from test point to line is minimum. And a digital example is given and the result is quite near to current method.
