Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/340630 
Year of Publication: 
2025
Citation: 
[Journal:] Borsa İstanbul Review [ISSN:] 2214-8469 [Volume:] 25 [Issue:] 6 [Year:] 2025 [Pages:] 1152-1165
Publisher: 
Elsevier, Amsterdam
Abstract: 
The dynamic field of financial markets is constantly in search of new ways to understand complex market dynamics. The increasing availability of vast amounts of text data offers new avenues for investigation (Botchway et al., 2020). This study aims to shed light on the dynamics between stock market movements and news narratives in Türkiye. To address this issue, the study will include the analysis of business, financial, and economic news from four major news journals (The Economist, The New York Times, The Guardian, and Yeni Şafak) along with local tweets. Yeni Şafak and local tweets serve as proxies for local news sentiment. The analysis rests on daily Turkish stock market data from January 1, 2015, to February 27, 2024, obtained from Yahoo Finance. The issue was addressed using state-of-the-art Natural Language Processing (NLP), machine learning, and explainable AI techniques. The findings reveal that international news significantly predicts the Turkish Stock market, with the majority of machine learning models yielding approximately 80 percent predictive accuracy. The Explainable AI methods demonstrate that traditional international news media have a significant impact on the Turkish stock market in comparison to local news sources such as Yeni Şafak and Twitter which serve as less effective predictors. Notably, the ensemble algorithms, comprising Random Forest, Gradient Boosting, and XGBoost, demonstrate robust performance across all datasets.
Subjects: 
Explainable AI
Machine learning
Natural language processing
Sentiment analysis
Stock market
Transformers
Persistent Identifier of the first edition: 
Creative Commons License: 
cc-by-nc-nd Logo
Document Type: 
Article
Appears in Collections:

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.