基于均差最小二乘拟合方程形式的研究  

Study on form of least squares fitting function Base on mean deviation

作　　者：高光发[1] 齐敏菊[2] 

机构地区：[1]安徽理工大学能源与安全学院 [2]淮南师范学院,安徽淮南232001

出　　处：《矿业工程》2010年第6期64-66,共3页Mining Engineering

基　　金：安徽省教育厅自然科学基金资助(2006kj009c);安徽理工大学青年科学基金资助;安徽理工大学科研启动基金资助(2006yb015)

摘　　要：分析了均差与拟合函数导数之间的关系,研究了拟合函数的形式及其中多项式最高次数的确定,以及拟合函数中系数的近似确定。结果表明,该方法可较好地确定拟合函数的具体形式,其确定的近似系数与实际值接近度较好。该方法对于提高科研拟合函数的精确度和减少拟合工作量具有借鉴意义。Relationship was analyzed between mean deviation and derivative of the fitting function deeply;with examples,study was carried out on the least-squares fitting-the form of fitting function and determination of the maximum number of polynomial,as well as the approximate coefficients of the fitting function.Fitting results indicate that this method can well determine the specific form of the fitting function;their approximate coefficients and the actual values are very close.The method is of guiding significance for improving the accuracy of fitting function and reducing the workload.

关 键 词：数据拟合 均差 最小二乘法 曲线拟合 

分 类 号：TB115[理学—数学]