Price prediction of dual-listed stocks with RF and LSTM algorithms: NYSE and BIST comparison
Year: 2024
Type: article
Abstract: Companies are looking for ways to access capital from developed markets instead of local markets to find financing. While some companies use debt instruments for this purpose, others use equity financing methods. One of the techniques used in equity financing is the simultaneous registration of shares on national and foreign stock exchanges, also known as the dual-registration method. Investors entering international markets by investing in dual-registered shares is important for companies to gain capital. However, another important issue for those investing in stocks is the ability to gain capital through accurate prediction of price movements. The aim of this study is to predict the prices of Turkcell stocks traded on Borsa Istanbul and the New York Stock Exchange (NYSE) using machine learning and deep learning methodologies. The results of the analyses conducted with the Random Forest Regressor and Long Short-Term Memory algorithms, which are machine learning and deep learning algorithms, respectively, showed that both algorithms exhibited a lower error rate in predicting the closing prices of Turkcell stocks on the NYSE.
Institutions: Necmettin Erbakan University, Selçuk University
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W4406334409