A comparison Between Principal Component Regression and Partial Least Squares Regression Methods with application in The Kirkuk Cement

Abstract

Appear in Many Application Areas for Regression Analysis and Presence the case of More Than One Variable Dependent Affected by A variety of  Explanatory Variable and at The Same Time The Number of Observation is Relatively Small Compared to The Number of  Variables, and Show Here The Problem of not Provide Offers Several Hypotheses Multiple Regression Analysis, More Over Prominence Problem of Multicollinearity between the Explanatory Variables Beside The Correlation between The Explanatory Variables and The Dependent Variables and Reflexive that on the Regression Estimates. and in This Research was Dealing with Problems from this Type Related to Variables Kirkuk Cement Factory, used the Methods, Principal Component PC and Partial Least Squares PLS to Solve the Problems Above, The First Method is Considered as one of the Commonest Methods used in Solving the Problem of Multicollinearity between the Explanatory Variables and the Second Method is Considered as one of  the Methods Which Dally Methodically Different in Deduction the Components Dependent on Curing The Correlation the Presence between The Explanatory Variables and The Dependent Variables. Through the Statistical Analysis, Orphan Conduction to The PLS Method it has Succeeded in Establishing the Optimal Regression Model for all Depended Variables, Besides Superiority This Method Whence Ability on the Prediction for Futuristic Values Apiece Dependent Variables and Also Whence Dimension Reduction. The (Minitab, Version, 16.1) is used in the Statistical Analysis for the Data of this Research .