在迴歸模型的應用上，迴歸不適用的模型中，有部份因素來自於其共線性，共線性主要產生的原因有二：
（1）來自於相似變數之間所產生的共線性
（2）一種是來自於複多項式模型假設所產生的共線性。
    本篇論文針對複多項式模型所產生的共線性提供標準步驟及sas程式來解決此問題。

In analyzing the regression model, the inadequate of model are sometimes due to the multi-collinearity among the independent variables. There are two major cases of this kind, one is the use of some similarity variables another one is fitting the multi-polynomial model. 
This thesis is concern of the case of using multi-polynomial model. 
We will provide some useful algorithms to establish a standard procedure in applying the ridge regression to analyze this kind of unacceptable result.

Cotents:

1.Introduction......1
2.Review of Multi-collinearity......1
  2.1 Some useful matrix properties......1
  2.2 Condition indices and Variance inflation factor......3
  2.3 Effect of multicollinearity......8
  2.4 Ridge Regression......13
3.Standard procedures for analyzing ridge regression......20
4.Reference......50
5.Appendix......50
  5.1 Programs for Paper......50
  5.2 SAS Macro......51
===================================
Table Cotents:

Table 2.2.1......4
Table 3.1	......20
Table 3.2......23
Table 3.3	......23
Table 3.4	......24
Table 3.5......26
Table 3.6......27
Table 3.7......28
Table 3.8......29
Table 3.9......30
Table 3.10......31
Table 3.11......32
Table 3.12......35
Table 3.13......36
Table 3.14......37
Table 3.15......37
Table 3.16......39
Table 3.17......41
Table 3.18......46
Table 3.19......47
Table 3.20......48
Table 3.21......49

===================================
Graph Cotents:

Graph 2.4.1......13
Graph 3.1......40
Graph 3.2......44

Hoerl & Kennard[1970a] “Ridge regression: Biased estimation for non-orthogonal roblems”,Technometrics,12,55-6

Kunugi, T., T. Tamura, and T. Naito [1961], “New acetylene process uses hydrogen dilution,” Chem. Eng. Prog., 57, 43-49