Long-Term General Index Prediction Based on Feature Selection and Search Methods: Amman Stock Exchange Market
Keywords:Amman stock index, feature selection and search methods, linear regression, economic sectors, prediction models, financial services, long-term general index, Syrian refugees, Syrian War, correlation analysis
Stock markets are an essential backbone for the economy worldwide; their indices provide all interested parties with indicators regarding the performance of firms listed in the financial market due to tracking the daily transactions. This study aims to investigate factors that affect the stock exchange directly so that it simplifies building a prediction model for the exchange index in Jordan’s financial market. The study hypothesis assumes that some sub-sectors are most influential in creating the stock market prediction model. Therefore, this study applies four feature selection methods on 23 sub-sectors and Amman Stock Exchange Index (ASEI100) for the period 2008–2018. The top 10 attributes from each selection method are combined, and the frequency table is used to find the highly trusted attributes. Moreover, linear regression with ordinary least square regression is used to test the validity of the top factors that frequently occurred in the four methods and their effect on ASEI. The results found that there are six main sub-sectors directly affecting the general index in Jordan: Health Care Services, Mining and Extraction Industries, Textiles, Leather and Clothing, Real Estate, Financial Services and Transportation. These sectors can be utilised to predict the movements of the Amman Stock Exchange Index in Jordan. Also, the linear regression model output showed a statistically significant relationship between the six sub-sectors (independent variables) and ASEI (dependent variable). Investors can use this paper’s findings to signal the most important sectors in Jordan. Thus, it helps in taking investment decisions.
How to Cite
This work is licensed under a Creative Commons Attribution 4.0 International License.