由一群資料尋找最適曲線是科學分支中一重要的工作。於本文中，我們研究最小平方法(Least squares algorithm)，辛普勒斯法(Simplex algorithm)及馬克特法(Marquardt's algorithm)並於IBM PC上實際比較其效率、精確度及其方法之適用性。

Fitting curves to data is an important task in various branches of science. In this paper, we investigate and implement (on an IBM-PC) the standard linear least squares algorithm, the simplex algorithm and Marquardt's algorithm. In particular, we shall compare the efficiency. accuracy, and general applicability of these algorithms.