Year: 2025
Type: article
Abstract: Stock market forecasting is the process of making future predictions based on past financial data, enabling individual and institutional investors to make informed investment decisions. One of the techniques widely used in forecasting studies, such as stock market index forecasting and stock price forecasting, is the Artificial Neural Network (ANN) method. This study aimed to forecast the BIST 100 index using data obtained from the Central Bank of the Republic of Turkey and other stock exchange websites between 02.01.2014 and 23.12.2024, using a multi-layered feedforward backpropagation ANN algorithm. Numerous model tests were conducted using BIST data for the period from 2014 to 2022. Within this scope, the best four models were determined based on statistical performance criteria, and the BIST 100 index forecast was analysed using data for the 2023-2024 period. As a result of the study, it was determined that the R² values of the other three models, except Model 2, were above the 60% success level; Model 1 reached 61.9%, Model 3 60.4% and Model 4 63.2% R² values, respectively.
Author: Hakan Yılmaz
Institution: Ağrı İbrahim Çeçen University
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4414536033