IMPROVEMENT THE PERFORMANCE OF THE SIMPLE LINEAR
REGRESSION MODEL BY USING THE WAVELET SHRINKAGE
Qais Mustafa Abdulqader
Dept.Banks Management, Zakho Technical Institute, Dohuk Polytechnic University, Zakho, Iraq
(Accepted for publication: October 30, 2014)
In this research, Two methods have been used to build a simple linear regression model through the application on real data and taking a simple random sample size of (128) observations. The first method depending on raw data, while the second method by using wavelet shrinkage .In the second method, the Daubechies wavelet (D4) was used to filter the raw data described in soft thresholding function with five multi-resolution levels of analysis and applying different methods to determine the levels of thresholding parameters. Results of the analysis showed the efficiency of wavelet shrinkage method in improving the performance of the linear regression model through reducing the value of the standard error of the estimate on the one hand and increasing the value of the coefficient of determination on the other hand compared with results of the estimated model of the raw data.
Keywords: linear Regression, Thresholding , Fixed Form , Minimax , Wavelet Shrinkage