索引表与逐步回归算法的改进与应用

对索引表算法进行了改进。主要是对原始数据文本按列循环读入一维数组之中，先对一维数组利用冒泡法进行排序，再利用索引表和最小二乘法对原始数据进行R估计，然后对拟合方程的自变量进行了高阶构造，利用主成分分析方法寻求主要影响因素，并结合最小二乘法，从而实现对逐步回归方法的替代。而后，根据以上计算原理建立了搜索最优方程的计算流程，采用这一计算流程可以查找异常值，然后利用EM算法对异常值进行了调整。

Improvement of Index Table and Successive Regression Algorithm and Their Application

The algorithm of index table is improved,the primordial data text is circularly read to a dimension array by columns,then being sorted by bubble sort algorithm.Every sorted datum is compared with a counted datum which is not among primordial data,and being circularly replaced and compared,so that the number of repeated data is counted in every dimension array and the rank of every element is computed by the definition of the rank,the primordial data can be used to R-estimation by finding index table and least squares method (LSE).The mainly effective independence is found by higher structure and the principal ingredient analysis method, and LES,successive regression method can be substituted.The computation flow for finding the best fitting equation is founded by LSE,M-estimation,R-estimation and successive regression,abnormal values can be found by the computation flow and adjusted by EM-algorithm.