Paper: arXiv 2506.06345
Authors: Sukru Selim Calik, Andac Akyuz, Zeynep Hilal Kilimci, Kerem Colak
Abstract
Financial literacy is increasingly dependent on the ability to interpret complex financial data and utilize advanced forecasting tools. In this context, this study proposes a novel approach that combines transformer-based time series models with explainable artificial intelligence (XAI) to enhance the interpretability and accuracy of stock price predictions. The analysis focuses on the daily stock prices of the five highest-volume banks listed in the BIST100 index, along with XBANK and XU100 indices, covering the period from January 2015 to March 2025. Models including DLinear, LTSNet, Vanilla Transformer, and Time Series Transformer are employed, with input features enriched by technical indicators. SHAP and LIME techniques are used to provide transparency into the influence of individual features on model outputs. The results demonstrate the strong predictive capabilities of transformer models and highlight the potential of interpretable machine learning to empower individuals in making informed investment decisions and actively engaging in financial markets.
Complexity vs Empirical Score
- Math Complexity: 7.5/10
- Empirical Rigor: 8.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced transformer architectures (DLinear, LTSNet, TST) and XAI techniques (SHAP/LIME) which involve dense mathematical foundations, while demonstrating strong empirical rigor with a decade-long backtest on BIST100 data using multiple models and technical indicators.
Research Flowchart
flowchart TD Start["Research Goal:<br>Predict stock prices with<br>high accuracy & interpretability"] --> Data["Data Source:<br>BIST100 (5 high-volume banks<br>XBANK & XU100 indices)<br>Jan 2015 - Mar 2025"] Data --> Features["Feature Engineering:<br>Technical Indicators<br>+ Raw Time Series Data"] Features --> Models["Computational Models:<br>DLinear, LTSNet,<br>Vanilla & Time Series Transformers"] Models --> XAI["Explainability Process:<br>SHAP & LIME analysis<br>on model predictions"] XAI --> Outcomes["Key Outcomes:<br>1. High prediction accuracy<br>2. Transparent feature influence<br>3. Actionable investment insights"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a novel approach that combines transformer-based time series models with explainable AI (XAI) to improve stock price prediction.
Source sentence
In this context, this study proposes a novel approach that combines transformer-based time series models with explainable artificial intelligence (XAI) to enhance the interpretability and accuracy of stock price predictions.
Assumes 3
- Only valid trading days on which the exchange is operational are included in the dataset. data
Source sentence
During the dataset construction process, only valid trading days on which the exchange is operational are included.
- Non‑trading days such as weekends and official holidays are systematically excluded. data
Source sentence
To maintain temporal consistency and ensure data integrity, non‑trading days such as weekends and official holidays are systematically excluded.
- DLinear assumes that time series behavior can be effectively modeled by applying linear operations to decomposed components. model
Source sentence
DLinear assumes that time series behavior can be effectively modeled by applying linear operations to decomposed components.
Data
- Turkish bank and index dataset market-prices public
Source sentence
... In addition to individual stock data, two principal Borsa Istanbul indices—XBANK (banking sector index) and XU100 (BIST 100 index)—are incorporated into the dataset
Costs
Not applicable
Finds 2
- DLinear consistently outperforms the other transformer models, achieving an R² of 0.9920 for ISCTR with a minimal RMSE of 0.3234. performance
Source sentence
DLinear achieves remarkably low error values (MSE, MAE, RMSE) and high R² scores, particularly in the case of ISCTR, where the R² value reaches 0.9920 with a minimal RMSE of 0.3234.
- The models explain between 94.7% and 99.3% of the variance in stock price movements across the evaluated assets. predictive-accuracy
Source sentence
cating that it explains between 94.7% and 99.3% of the variance in stock price movements across the evaluated assets.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 18 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.