Theses and Dissertations - UTB/UTPA
Date of Award
Master of Science (MS)
Dr. Douglas Timmer
Dr. Miguel A. Gonzalez
Dr. Alley Butler
There are many instances in industry, where the necessity of prediction models is extremely important. However, the difficulty of forecasting increases with the number of variables involved, and determining independent variables from dependent variables. Autocorrelation is one of the most common causes for many control charts and other quality measurement tools to falsely signal an error in the process. Time series models, such as autoregressive models, accommodate autocorrelated data to a certain extent. However, using a model that only analyzes historical data alone does not possess the accuracy necessary to be able to forecast data precisely. An addition of an independent variable into an AR(1) model will increase the accuracy by noting how correlated the two variables are and using that value in order to correct the previous errors. Likelihood functions will aid in the derivation of the model. The results will develop a new robust control chart.
University of Texas-Pan American
Copyright 2012 Becky M. Vela. All rights reserved.