EUVIMEDEUVIMED entdecken
Systematische ÜbersichtOpen Access · CC BY 4.0

Predictive analytics and NLP for financial forecasting: a systematic review of advances, ethical challenges, and future directions

Nurudeen Jibrin, Abdulhamid Abubakar, Mohammed Umar Iya, Gilbert I. Aimufua, Monday O. Adenomon

Journal of Statistical Sciences and Computational Intelligence · 15.06.2026 · 2 (2) · 576-600

DOI: 10.64497/jssci.98 ↗

Kurz gefasst

Systematische Übersicht in der Zeitschrift Journal of Statistical Sciences and Computational Intelligence, online erschienen am 15.06.2026. Verfasst von Nurudeen Jibrin und 4 weiteren Autor:innen. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Predictive analytics and natural language processing are reshaping financial forecasting by enabling data-driven market intelligence. However, persistent challenges including model interpretability, ethical risk, and the integration of hybrid architectures limit their practical deployment. This study synthesises advancements, performance trade-offs, and ethical considerations in this rapidly evolving field. A systematic literature review was conducted following PRISMA 2020 guidelines. Searches across IEEE Xplore, ScienceDirect, SpringerLink, Scopus, Web of Science, ACM Digital Library, arXiv, SSRN, and Google Scholar yielded 1,856 records. After deduplication and screening, 50 studies published between 2020 and 2026 were included. A standardised extraction form captured methodological details, performance metrics, interpretability features, computational costs, and ethical dimensions. Comparative analysis benchmarked four model classes: ARIMA, LSTM networks, FinBERT, and hybrid ARIMA-LSTM architectures. Hybrid ARIMA-LSTM models achieve the highest normalised accuracy (0.91), balancing predictive performance with improved interpretability over standalone LSTM, albeit at higher computational cost. LSTM excels during volatile market periods (0.85), while FinBERT dominates sentiment-driven contexts (0.81). Ethical risk profiling reveals systemic bias vulnerabilities across five documented categories (look-ahead, survivorship, narrative, objective, and cost bias), yet no single bias is discussed in more than 28 per cent of studies. Temporal adoption mapping indicates a cautious integration cycle within financial institutions, with transformers emerging between 2020 and 2025. The study proposes a five-year research roadmap prioritising interpretable hybrid modelling (2026 - 2027), multi-modal fusion protocols (2027-2028), and regulatory-grade explainable artificial intelligence (2028 - 2030). These coordinated advancements promise to balance predictive precision with ethical accountability in next-generation financial intelligence.

Zitieren

APA

Jibrin, N., Abubakar, A., Iya, M. U., Aimufua, G. I., & Adenomon, M. O. (2026). Predictive analytics and NLP for financial forecasting: a systematic review of advances, ethical challenges, and future directions. Journal of Statistical Sciences and Computational Intelligence, 2(2), 576-600. https://doi.org/10.64497/jssci.98

Vancouver

Jibrin N, Abubakar A, Iya MU, Aimufua GI, Adenomon MO. Predictive analytics and NLP for financial forecasting: a systematic review of advances, ethical challenges, and future directions. Journal of Statistical Sciences and Computational Intelligence. 2026;2(2):576-600. doi:10.64497/jssci.98

BibTeX
@article{Jibrin2026,
  author = {Jibrin, Nurudeen and Abubakar, Abdulhamid and Iya, Mohammed Umar and Aimufua, Gilbert I. and Adenomon, Monday O.},
  title = {Predictive analytics and NLP for financial forecasting: a systematic review of advances, ethical challenges, and future directions},
  journal = {Journal of Statistical Sciences and Computational Intelligence},
  year = {2026},
  volume = {2},
  number = {2},
  pages = {576--600},
  doi = {10.64497/jssci.98}
}
Kostenlos starten →