College of Computer Science and Mathematics

Discussion of the master’s thesis in the College of Computer Science and Mathematics -Department of Statistics and Informatics Department entitled:

Hybridizing Bayesian ARIMA and Echo State Network for Chaotic Time Series Forecasting

Master’s thesis by  Shahla Tahseen Hasan Zakar

supervised by

Assis.Prof.Dr. Osamah Basheer Shukur Mahmood

 

This thesis includes:

The thesis dealt with the development of a hybrid model for predicting chaotic time series, by combining Bayesian analysis of the ARIMA model with an echo-state network (ESN).

 

The study addressed:

The study explored the integration of Bayesian analysis to address uncertainty with ESN networks to capture nonlinear patterns. It also seasonally separated the data (hot and cold) to increase accuracy, thus overcoming the limitations of traditional models.

 

 

The study aims:

This study aims to improve the accuracy of forecasting complex weather phenomena through an approach that combines Bayesian analysis to address uncertainty in parameter estimation, ensuring higher accuracy and greater reliability in forecasts, with the nonlinear capabilities of an echo-state network (ESN). It presents a hybrid predictive model (Bayesian AR(p)-ESN) to enhance the forecast quality of random time series data for daily wind speed in Mosul.

 

The discussion committee consists of:

 

Dr. Muthanna Subhi Sulaiman (Chairman)

Dr. Rikan Abdul Aziz Ahmed (Member)

Dr. Omar Salem Ibrahim (Member)

Dr. Osamah Basheer Shukur (Member and Supervisor)