Using Multilevel Poisson Regression with Wavelets Filters

Othman Babakr, Sawen; AbdulMajeed Badal, Mohammed

	Using Multilevel Poisson Regression with Wavelets Filters
Journal of Kirkuk University For Administrative and Economic Sciences
Volume 13, Issue 1, June 2023, Pages 179-192 PDF (1.35 M)
Authors
Sawen Othman babakr; Mohammed AbdulMajeed Badal
University of Kirkuk Journal For Administrative and Economic Science
Abstract
The proposed method in this study is used to identify the influence of the drug on the thyroid disease as well as the ordinary least squares method (OLS) and the generalized linear regression method (GLM) are used to prepare the estimate of the linear model. The study concluded that wavelet filters (Haar, Symlets, and Daubechies) produce the best results for estimating the Poisson Regression model when compared to the Generalized Linear Regression model based on (MSE, R2, and P-value). The data in this study has been collected from Smart Health Tower in Sulaimani, which contains 128 cases of thyroid disease, as well as the simulation data used in this study. The paper's finding showed that the STOP-TSH (sym72) wavelet with the (fixed from thresholding) threshold technique and soft threshold rule was the most efficient when compared to all other proposed methods and the classical approach for both real and simulation data. The STOP_TSH (sym52) wavelet with the (fixed from thresholding) threshold method and hard threshold rule was the most efficient compared with all other proposed methods and the classical method for both real and simulation data. All the proposed methods have better efficiency than the classical method in estimating the Poisson Regression model, depending on the average of Akaike Information Criteria (AIC) and Bayesian information criteria (BIC).
Keywords
Poisson Regression; GLM; Wavelet Shrinkage; thresholding rules

Statistics Article View: 3,269 PDF Download: 231