Selection Of Difference Equation Minimum Model Specification By Correlation Analysis

Mohammad Hafiz, Abdellah (2010) Selection Of Difference Equation Minimum Model Specification By Correlation Analysis. Project Report. UTeM, Melaka, Malaysia. (Submitted)

	PDF (24 pages) Selection_Of_Difference_Equation_Minimum_Model_Specification_By_Correlation_Analysis_24_pages.pdf - Submitted Version Download (262kB)
	PDF (Full Text) Selection_Of_Difference_Equation_Minimum_Model_Specification_By_Correlation_Analysis_Full_Text.pdf - Submitted Version Restricted to Registered users only Download (779kB)

Abstract

This project would give more understanding about least square method, model structure selection, difference equation model and correlation test. Through the study, the significance of correlation analysis in selection of difference equation minimum model specification is understood and the knowledge of variable correlation analysis is utilized in selection of difference equation minimum model specification. The scopes of the project are limited to implement the study to five simulated linear and nonlinear difference equation model. The mathematical models that apply in this research are ARX and NARX model. In system identification both the determination of model structure and model validation are important steps. Autocorrelation test and crosscorrelations are two of the correlation analyses used. The results were obtained through programming using Matlab software and then compared in order to identify and select the difference equation minimum model specification by correlation analysis. It is shown that the application and interpretation of these tests to simulated data were able to identify a simpler if not the same, model specification as the original specification.

Actions (login required)

View Item