A study at the University of Karbala discusses Forecasting Common Stock Returns Using Radial Neural Networks An application Study in Iraqi Stock Exchange for the period (2041-2023)

You are currently viewing A study at the University of Karbala discusses Forecasting Common Stock Returns Using Radial Neural Networks An application Study in Iraqi Stock Exchange for the period (2041-2023)

A study at the University of Karbala discusses Forecasting Common Stock Returns Using Radial Neural Networks An application Study in Iraqi Stock Exchange for the period (2041-2023)

Forecasting Common Stock Returns Using Radial Neural Networks An application Study in Iraqi Stock Exchange for the period (2041-2023)

A Thesis Submitted To The Council of College of Administration and
Economics at Karbala University, As Partial Fulfillment of the
Requirements for PH.D. Degree Financial and Banking Sciences.

Submitted by
Hussam Kamel Sultan

Under the supervision

Professor Dr. Haider Younis Al-Moussawi
Assistant Professor Dr. Haider Abbas Al-Janabi

The study aims to improve equity investment decisions by Forecasting their returns using a series of historical data for a sample of companies listed on the Iraq Stock Exchange. The study also seeks to test proposed prediction models,
including the Box-Jenkins models and radial neural networks, with the aim of arriving at the best and most accurate model for investors to use in making their investment decisions. Radial neural networks are a subject of intellectual
and applied controversy, as many question the validity and superiority of these models in Forecasting equity returns, especially after the increased debate about the superiority of one over the other. Therefore, this study aims to shed
light on this issue and attempt to resolve it by testing the aforementioned models based on data collected for the study sample, represented by the Iraq Stock Exchange, and using a set of financial, statistical, and mathematical
methods. The study population is represented by the sectors and companies listed on the Iraq Stock Exchange, while the study sample is limited to (7) sectors by taking one model from each sector, meaning the number of companies is (7) companies for a period of (10) years. (January 1, 2014) to December 31, 2023) on a weekly basis, with four observations per month. The study reached several conclusions, the most important of which is that the radial basis function (RBF) network model provided predictions close to the actual price values, which indicates the quality of the predictions provided by this method. The study also came up with a number of recommendations: Investors should direct their wealth towards selecting profitable stocks that contribute to maximizing their wealth. To achieve this, it is recommended to adopt RBF neural network models, as these models help in making more effective investment decisions, as well as reducing time, effort and cost, and improving the quality of financial services.