对来自中文源文献的64张正交表按其有无平方和、自由度的自检功能进行了分组归类。选定其中19张正交表为C语言编程的内核，将套每张正交表进行方差分析的算法编写成19个可由主函数调用的子程序。在此过程中添加“自动检校”程序，以套L9(34)分析三因素正交试验结果的编程运行为例，首次对算法的功能正确性通过C程序运行结果同时显示出来。再借助Windows2000平台的Turbo C 2.0集成环境，形成可供150种以上的复因素水平组合方案按经典试验设计的数据结构进行方差分析的通用程序源文件，为不依赖统计软件而进行的数据挖掘工作拓展了新的路径。

The 64 orthogonal arrays, which originated from the literatures in Chinese, will be categorized according to self-checking of quadratic sum and degree of freedom. The 19 orthogonal arrays in 64 will be selected as the inner core of C program. The algorithms in which the variance of each orthogonal array is analyzed will be designed as 19 subprograms. In this process, the self-checking program will be added, with an example of the running and programming for analysis on results of tri-factors-orthogonal trials by L9(34), the correctness of algorithms will be showed with the running results of C in the same time. Used the Turbo C in Windows 2000, the general source program files, by which more than 150 multi-factor combined schemes are analyzed by data structure’s variance according to the design of classical trials, will be formed. It will establish a new method for data mining but not depend on statistical software.