Skip to content
Log in
Sponsored Artificial Intelligence: A Modern Approach, Global Edition Artificial Intelligence: A Modern Approach, Global Edition
Review Open access

Explainable AI in Financial Forecasting: Balancing Predictive Accuracy and Economic Interpretability

Taib Ali Sumera Begum Hassan Khan Muhammad Saif Ul Islam Rao
Aug 2026 · Journal of Business Insight and Innovation · Vol 5, pp. 86-95 · 0 citations · 22 references

TL;DR

None of the models in the study achieve both the highest accuracy and interpretability in financial forecasting models, but an ensemble model was found to be a good compromise between the trade-off of prediction accuracy and interpretability.

Abstract

Background: Financial forecasting has increasingly turned to artificial intelligence (AI) techniques, which can help boost the accuracy of forecasts in complex and volatile markets. However, many advanced AI models are "black boxes," lacking transparency and invoking concerns regarding interpretability, trust, and regulatory compliance of AI-powered solutions. In the world of financial decision-making systems, the challenge lies in balancing high predictive capabilities with economic interpretability. The issue of balancing high predictive capabilities with economic interpretability has emerged in the world of financial decision-making systems as a solution, namely, Explainable AI (XAI). Objective: This study seeks to test the performance of XAI methods to strike a balance between predictive performance and economic interpretability in financial forecasting models. Methods: The method used in this study was a quantitative comparative research design where the financial time series data were multivariate data consisting of stock prices, inflation rates, interest rates, trading volumes, and GDP growth. Various models such as ARIMA, linear regression, random forest, gradient boosting, and LSTM networks were implemented. The performance of the models was tested using the following metrics: MSE, RMSE, MAE, and R². Model interpretation was performed with XAI techniques (SHAP, LIME) and feature importance, fidelity, and stability measures were used to evaluate interpretability. Results: Deep learning models, specifically LSTM models, had the best predictive accuracy with R² = 0.94, with linear models having the best interpretability score. The most important predictors were identified by SHAP analysis, which showed interest rates and inflation. An ensemble model was found to be a good compromise between the trade-off of prediction accuracy and interpretability. Conclusion: None of the models in the study achieve both the highest accuracy and interpretability. In financial applications with the demands of trust and compliance in mind, complex financial models can be significantly improved in terms of transparency using XAI techniques. References Adenekan, T. K. (2025). Explainable AI in financial decision-making systems. Journal of Financial Data Science, 7(2), 45–62. https://doi.org/10.2139/ssrn.5123456  Adinya, I., & Udomboso, C. (2025). Modeling and forecasting of financial time series in emerging markets using multilayer perceptrons. University of Ibadan Journal of Science and Logics in ICT Research, 14(1), 78–95. Arsenault, P. D., Wang, S., & Patenaude, J. M. (2024). Explainable artificial intelligence in financial forecasting: A review. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.4876543 Arsenault, P.-D., Wang, S., & Patenaude, J.-M. (2025). A survey of explainable artificial intelligence (XAI) in financial time series forecasting. ACM Computing Surveys, 57(10), Article 265, 1–37. https://doi.org/10.1145/3729531 Bank of England. (2024). Artificial intelligence and machine learning in financial services (Bank of England Quarterly Bulletin, 2024 Q3). Bank of England. Bank of England. (2024). Artificial intelligence in UK financial services (Financial Stability Report, November 2024). Bank of England. Carta, S. (2021). Explainable AI for financial forecasting. In Proceedings of the 2021 International Joint Conference on Neural Networks (IJCNN) (pp. 1–8). IEEE. https://doi.org/10.1109/IJCNN52387.2021.9534098  Çelik, T. B., İcan, Ö., & Bulut, E. (2023). Extending machine learning prediction capabilities by explainable AI in financial time series prediction. Applied Soft Computing, 132, 109876. https://doi.org/10.1016/j.asoc.2022.109876 Černevičienė, J., & Kabašinskas, A. (2024). Explainable artificial intelligence in finance: A systematic review. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.4987654  Chen, X. Q. (2023). Explainable artificial intelligence in finance: A bibliometric analysis. Journal of Business Research, 170, 114321. https://doi.org/10.1016/j.jbusres.2023.114321  Chen, X. Q. (2023). Explainable artificial intelligence in finance: A review of methodologies. International Journal of Finance & Economics, 28(4), 4567–4589. https://doi.org/10.1002/ijfe.2789  Černevičienė, J., & Kabašinskas, A. (2024). Explainable AI in financial forecasting systems. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.4987655  Cheng, L. (2024). Integrating explainable AI techniques in time series price forecasting for enhanced interpretability and decision-making (SSRN Working Paper No. 5373836). SSRN. https://doi.org/10.2139/ssrn.5373836 Giudici, P. (2024). Explainable AI methods for financial time series. Journal of Financial Econometrics, 22(3), 621–645. https://doi.org/10.1093/jjfinec/nbae012  Giudici, P. (2024). Machine learning and explainability in financial markets. Quantitative Finance, 24(6), 845–862. https://doi.org/10.1080/14697688.2024.2345678  Gupta, S. (2023). AI-driven financial forecasting and enterprise transformation. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.4765432  Gupta, Y., Naresh, P., Nagesh, C., Praveen, P., Kirankumar, N., & Surendra, K. (2025, July). An interpretable framework for stock market forecasting using long short-term memory (LSTM) networks with SHAP-explainable artificial intelligence (XAI) method. In Proceedings of the 2025 6th International Conference on Data Intelligence and Cognitive Informatics (ICDICI) (pp. 539–546). IEEE. Iqbal, A., & Amin, R. (2025). An efficient mechanism for time series forecasting and anomaly detection using explainable artificial intelligence. The Journal of Supercomputing, 81(4), 523. https://doi.org/10.1007/s11227-025-07128-8 Isufaj, A. (2025). Towards explainable and reliable AI in finance. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.5109876  Isufaj, A. (2025). Trustworthy AI systems in financial applications. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.5109877  Kumar, S., Kumar, P., Sidhu, R., Kumari, N., Kaur, S., & Singh, K. N. (2025, November). Empirical analysis of AI methods of forecasting financial time series: A research on LSTM and ML models in Indian stock, commodity and energy prices. In Proceedings of the 2025 International Conference on Innovations and Emerging Technologies in AI & Communication Systems (IETACS) (pp. 800–805). IEEE. Liu, R., Yi, S., Ablikim, A., & Song, X. (2025). A MF-ConvLSTM-XAI model integrating multi-feature and fuzzy control for financial time series forecasting. Scientific Reports, 15, Article 12345. https://doi.org/10.1038/s41598-025-98462-8  Maarif, M. R., Saleh, A. R., Habibi, M., Fitriyani, N. L., & Syafrudin, M. (2023). Energy usage forecasting model based on long short-term memory (LSTM) and explainable artificial intelligence (XAI). Information, 14(5), 265. https://doi.org/10.3390/info14050265 Marey, N., Abu-Musa, A. A., & Ganna, M. (2024). Integrating deep learning and explainable artificial intelligence techniques for stock price predictions: An empirical study based on time series big data. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.4987656  Mohsin, M. T., & Nasim, N. B. (2025). Explainable AI for financial prediction: A systematic review. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.5123460  Pantov, B. (2024). Explainable AI applications in financial systems. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.4987657  Ridwan, L. (2025). Explainable artificial intelligence (XAI) in financial forecasting: Transparency and trust in stock market predictions. Journal of Financial Technology, 11(2), 112–130. https://doi.org/10.1080/23322039.2025.2456789  Ridwan, L. (2025). XAI in financial forecasting: Trust and transparency. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.5123461  Singh, J. (2026). The new frontier of financial forecasting: A comprehensive review of deep learning architectures, hybrid models, and the imperative of explainable AI in global stock markets. AIJFR-Advanced International Journal for Research, 7(1), 1–25. Sodnomdavaa, T. (2026). Explainable and hybrid AI approaches for corporate financial performance forecasting: A structured literature review. International Journal of Accounting Information Systems, 58, 100687. https://doi.org/10.1016/j.accinf.2026.100687  Sodnomdavaa, T. (2026). Hybrid explainable AI approaches for financial forecasting. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.5123462  Sodnomdavaa, T., & Bataa, A. (2025). Explainable hybrid modeling of stock market turning points: An integrated ARIMA–LSTM–SHAP approach. [Working paper], Available at SSRN. https://doi.org/10.2139/ssrn.5123463  Tanveer, H. (2024). Explainable AI in stock market forecasting. Journal of Forecasting, 43(6), 1987–2005. https://doi.org/10.1002/for.3123 

DOI Semantic Scholar
Read PDF

Similar papers

Review Open access Aug 2026

Artificial Intelligence in Financial Decision-Making: Opportunities, Challenges and Implications for the Modern Financial Sector

Artificial Intelligence (AI) is quickly taking center stage in modern financial decision-making. The capacity of financial institutions and investors to analyze vast and intricate data sets has increased thanks to developments in machine learning, deep learning, natural language processing, predictive analytics, and generative artificial intelligence. Conventional methods for investment analysis, credit evaluation, fraud detection, financial forecasting, risk assessment, and portfolio management have been altered by the growing use of AI. This study looks at the expanding use of AI in financial decision-making and assesses both its possible advantages and the difficulties in implementing it. While taking into account issues with data quality, algorithmic bias, privacy, cybersecurity, explainability, model risk, and an over-reliance on automated systems, the study focuses on how AI can enhance the speed, consistency, analytical depth, and efficiency of financial decisions. Using contemporary scholarly and institutional literature on AI and finance, a descriptive and analytical research technique is used. According to the investigation, AI may greatly improve financial decision-support by digesting data quickly and seeing connections that conventional analytical techniques can miss. However, data quality, model architecture, governance, and human oversight all have a significant impact on how successful AI is. Increasing the use of AI can potentially lead to new vulnerabilities in the financial industry, especially through third-party concentration, cyber risks, market correlations, and model risk, according to recent international data. The study comes to the conclusion that rather than completely replacing human judgment, AI should be included into finance largely as an enhancement of human competence. Strong governance, open decision-making procedures, trustworthy data, ongoing model review, and significant human monitoring are all necessary for responsible deployment. The study offers a theoretical framework for comprehending how financial institutions might profit from AI while managing the dangers related to technology, ethics, and finances.

Shalu, Garima, Bhumika, Dr. Bhawana · 0 citations
Jul 2026

Role of artificial intelligence in financial decision-making among investors: A quantitative study

The study aims to examine the impact of artificial intelligence (AI) on the quality of financial decision-making of investors in Maharashtra, India. Specifically, it examines the impact of AI predictive analytics (AIPA), data processing speed (DPS), bias reduction (BR), and cost savings (CS) on financial decision-making quality (FDMQ) and the moderating effect of investor experience (IE). Data was gathered from 228 investors in Maharashtra through a closed-ended questionnaire administered through Google Forms and WhatsApp. Structural Equation Modeling (SEM) was employed using SmartPLS 4 to examine the hypothesized relations. The research finds that AIPA, DPS, BR, and CS considerably enhance FDMQ. Moreover, investor experience positively moderates these connections and strengthens the role of AI on decision quality. This research confirms the utility of incorporating AI into investment decisions aimed at enhanced precision, speed, and justice of decisions. Moreover, experience-led training and tactful usage of AI aids become necessary to benefit from various profiles of investors. This study contributes to the current body of research on AI in finance and provides actionable recommendations for investors, financial institutions, and policymakers with empirical evidence from one of India's leading financial hubs.

C. Anirvinna, Jyoti Ranjana, Babita Jha et al. · 0 citations
Preprint Aug 2026

AI-Driven Multiscenario Interest Rate Forecasting: A Proof of Concept for Banking Asset Management

This study focuses on developing an AI-supported prototype for multiperspective interest rate forecasting that combines classical econometric models with modern artificial intel-ligence methods. Tested in a major European bank, the system enables more precise and flexible prediction of interest rate developments, supporting strategic decision-making in Asset-Liability Management (ALM). It integrates topic modeling, sentiment analysis, econometric forecasting, and market-based analyses within an interactive platform. Leveraging AI to analyze large volumes of financial documents and market data enables the identification of monetary policy trends and sentiment signals at an early stage. The core econometric model is a Bayesian vector autoregression (BVAR) that enables simulation-based scenario analyses to evaluate economic developments from multiple perspectives. The system's innovation lies in its integration of several forecasting approaches that consolidate previously separate information sources and present them transparently and interpretably. Financial analysts and risk managers thus gain a better basis for making decisions, allowing them to assess interest rate risks more accurately and manage market movements more proactively. While the prototype demonstrates how AI can transform interest rate management in banking, further development is required to optimize real-time data integration and regulatory compliance. Even at this stage, the study shows that multi-perspective, AI-driven forecasting provides substantial added value for banks by increasing transparency, strengthening evidence-based decision-making, and improving risk management.

Ekkehardt Bauer, Dirk A. Hollander, David Scholz et al. · 0 citations
Open access Jul 2026

Beyond Forecast Accuracy: Evaluating the Error–Profit Paradox in AI-Based Copper Price Prediction

Copper is a strategically important commodity whose price dynamics are increasingly affected by structural changes, geopolitical shocks, and the global energy transition. These conditions create substantial challenges for forecasting models and provide a useful setting for evaluating the practical value of machine learning predictions. This study compares statistical and artificial intelligence-based forecasting models for copper price prediction under different market regimes and structural break conditions. Model performance is assessed using a multi-dimensional evaluation framework that combines statistical accuracy (MAPE), dynamic pattern reproduction (Taylor diagrams and time-lagged cross-correlation analysis), and the economic performance of forecast-driven trading strategies. The results reveal a consistent error–profit paradox: models with the highest statistical forecasting accuracy do not necessarily generate the best trading outcomes. In several cases, models with larger prediction errors achieve superior economic performance because they capture directional market dynamics more effectively. The analyses further show that structural breaks substantially alter model rankings and predictive usefulness, highlighting the importance of regime-aware evaluation. These findings suggest that forecast accuracy alone provides an incomplete assessment of model quality in financial and commodity forecasting applications. The study contributes to machine learning evaluation research by proposing an integrated framework that jointly considers predictive accuracy, temporal dynamics, model robustness, and economic utility, thereby offering a more comprehensive approach to assessing forecasting systems in real-world decision-making environments.

László Vancsura, Tibor Tatay, Tibor Bareith · 0 citations
Review Open access Aug 2026

Artificial Intelligence in Predictive Analytics for Financial Forecasting

The application of Artificial Intelligence (AI) in predictive analytics has revolutionized the field of financial forecasting. This paper explores the impact of AI on financial forecasting by examining its methodologies, benefits, challenges, and future directions. By leveraging machine learning, neural networks, and advanced statistical models, AI enhances the accuracy and efficiency of financial predictions, aiding businesses in strategic planning and risk management. This study adopted an exploratory research design, which is particularly suitable for investigating phenomena where limited prior research exists or where the researcher seeks to gain deeper insights into the subject matter. The exploratory approach enabled the researcher to examine the relationship between non-payment of salaries and employee productivity in a flexible and comprehensive manner. The exploratory design was operationalized through the systematic review of existing literature and empirical studies. This involved gathering and analyzing secondary data from published journal articles, books, theses, dissertations, organizational reports, and other relevant academic sources. By reviewing existing studies, the researcher was able to identify patterns, themes, and relationships that have been established by previous researchers regarding salary payment issues and their effects on employee productivity. The literature review highlights the rapid growth of AI in predictive analytics for financial forecasting, covering key areas such as financial management, credit risk analysis, portfolio management, and fraud detection. AI algorithms, particularly those utilizing machine learning and neural networks, significantly improve forecast accuracy by capturing intricate patterns and relationships that traditional methods may overlook. By handling large datasets and employing advanced modeling techniques, AI automates forecasting processes, reducing analysis time and effort. This automation facilitates more frequent and scalable forecasting, enabling agile decision-making in businesses. Moreover, AI's capability to integrate diverse data sources, including structured financial data and unstructured text, offers a comprehensive perspective on market trends and risks, thereby enhancing the reliability of financial forecasts.

Muyiwa Emmanuel Dagunduro · 0 citations
Open access Jul 2026

TRuE-XAI: causal and explainable ai framework for trustworthy corporate earnings growth forecasting

This study proposes TRuE-XAI (Transparent, Rule-based, and Explainable Artificial Intelligence), an integrated framework combining imbalance-aware ensemble learning, automated hyperparameter optimization, rule-based explainability, visual analytics, and causal inference for transparent earnings-growth forecasting.

G. Jamnal · 0 citations