翻訳と辞書 |
Quasi-Newton Least Squares Method : ウィキペディア英語版 | Quasi-Newton Least Squares Method
In numerical analysis, The Quasi-Newton Least Squares Method is a quasi-Newton method for finding roots in variables. It was originally described by Rob Haelterman et al. in 2009. Newton's method for solving uses the Jacobian matrix, , at every iteration. However, computing this Jacobian is a difficult (sometimes even impossible) and expensive operation. The idea behind the Quasi-Newton Least Squares Method is to build up an approximate Jacobian based on known input-output pairs of the function . Haelterman et al. also showed that when the Quasi-Newton Least Squares Method is applied to a linear system of size , it converges in at most steps although like all quasi-Newton methods, it may not converge for nonlinear systems. The method is closely related to the Quasi-Newton Inverse Least Squares Method. ==References==
抄文引用元・出典: フリー百科事典『 ウィキペディア(Wikipedia)』 ■ウィキペディアで「Quasi-Newton Least Squares Method」の詳細全文を読む
スポンサード リンク
翻訳と辞書 : 翻訳のためのインターネットリソース |
Copyright(C) kotoba.ne.jp 1997-2016. All Rights Reserved.
|
|