最小2乗法における不良データの影響
スポンサーリンク
概要
- 論文の詳細を見る
The main purpose is the systematic study of the influences of inappropriate data in the method of least squares upon its results, where the inappropriate data mean not only the observation error but the unexpected error, for example, reading error at data producing, or punching error at input data of computer, and so on, which are not considered as the Ergodic time series. The influences of inappropriate data are researched for each as follow; 1. a coefficent or observed value of observation equations at indirect observation, 2. a coefficent or observed value of condition equations at direct observation. For this purpose, the analysis has been done through convenient formulas by matrices representation, and the efforts have become very economized for repeat computation. Moreover, it is indicated that it is able to find out the positions of inappropriate data by calucuiation of the residuals in the method of the least squares. Some examples are attached for application of this method.
- 一般社団法人情報処理学会の論文
- 1968-05-15
著者
関連論文
- 中川 徹, 小柳義夫 著, "最小二乗法による実験データ解析", 東京大学出版会, A5判, 206p., \2,400, 1982
- 最小2乗法における不良データの影響
- 連立一次方程式における係数の変化と解の変化