معالجة مشكلة التداخل الخطي باستخدام تحليل المكونات الرئيسية : بالتطبيق على استهلاك الوقود بالسيارات

Abstract EN

This study aimed at addressing the issues of Multicollinearity among explanatory variables using principal components analysis, eleven variables representing the mechanical and physical characteristics of the vehicles have been selected, in addition one more dependent variable representing gasoline consumption, data in this study taken from:(David J.

Balding, Noel A.

C and other, (200 )6 "Regression Analysis by Example Fourth Edition"Inc., Hoboken, New Jersey Published simultaneously in Canada).Hypotheses of the study are as follows: General model of the data can solve the problem of multicollinearity among the explanatory variables.

The regression model provides better result than the result obtained through the other estimations method.

Principal components regression method leads to removal of colllinearity among explanatory variables.

The SPSS program has been use in this is study for the purpose of analyzing data, where the explanatory variables have been found to be undergoing a problem of multicollinearity, and the problem has been solved through the principal component analysis.

Most important outcomes of the study include: The principal component analysis is an effective method for detection of multicollinearity, for the reason of that the principal component at always orthogonal.

The regression model obtained through the principal component is somewhat successful, has less variance and can be relayed on when it comes to prediction.

The PCR is more effective than the OLS since has less MSE value.

There are several criteria standards for detecting multicollinearity and problem solution, the principal component regression one of those standards, in where of that it can be easily understood and applied.