مقارنة مقدرات عرض الحزمة (معلمة التمهيد) باستخدام الدوال اللبية في تحليل المركبات الرئيسية

Abstract

Always Principal Component Analysis (PCA) used multivariate analysis with high dimensional data sets, often using Principal Component Analysis (PCA) to reduce these dimensions, The Principal Component Analysis (PCA) based on the study of the relationship between a group of high-dimensional variables and convert them to a new group of components,The principle component analysis based on study the relation between the high dimension variable group and transfer it to new groups of components Which summarizes the measured variables and be qualified to explain the most of the contrast of the original data , and it depends mainly on the calculation of the covariance matrix or correlation matrix but this analysis is affected by the nature of the data, it can be performed only after the covariance matrix achieve to the conditions before starting their own analysis. One of these condition is the linearity of data , Since the data of the phenomenon studied in this research do not achieve this status being characterized by non-linear feature, so it was resorting to the use of Kernel functions in the analysis of the principle components ,which is based on its calculated on the beginning identification (bandwidth) estimated in our research with variety methods ,are Least Squares Cross Validation (LSCV), biased crossing valid (BCV), Smoothed Cross-Validation (SCV) , Direct Plug-in Rule (DPI) . comparing these estimatore through the effective principle components numbers when their eigen values increasing more than one and corresponds from the ratio in explain the Total variance .