A RISK-AWARE HYBRID ENSEMBLE APPROACH FOR AEX INDEX FORECASTING: INTEGRATING APARCH-T VOLATILITY WITH LSTM-CNN-RF ARCHITECTURES

Aman Shreevastava, Bharat Kumar Meher, Ramona Birau, Virgil Popescu, Shahil Raza, Gabriela Ana Maria Lupu (filip) and Stefan Margaritescu
Additional contact information
Aman Shreevastava: P.G. DEPARTMENT OF COMMERCE AND MANAGEMENT, PURNEA UNIVERSITY, PURNEA, BIHAR, INDIA-854301
Bharat Kumar Meher: P.G. DEPARTMENT OF COMMERCE AND MANAGEMENT, PURNEA UNIVERSITY, PURNEA, BIHAR, INDIA-854301
Ramona Birau: UNIVERSITY OF CRAIOVA, "EUGENIU CARADA" DOCTORAL SCHOOL OF ECONOMIC SCIENCES, CRAIOVA, ROMANIA & CONSTANTIN BRANCUSI UNIVERSITY OF TARGU JIU, FACULTY OF ECONOMIC SCIENCE, TG-JIU, ROMANIA
Virgil Popescu: FACULTY OF ECONOMICS AND BUSINESS ADMINISTRATION, UNIVERSITY OF CRAIOVA, CRAIOVA, ROMANIA
Shahil Raza: DEPARTMENT OF COMMERCE, ALIGARH MUSLIM UNIVERSITY, ALIGARH, UTTAR PRADESH 202001, INDIA
Gabriela Ana Maria Lupu (filip): UNIVERSITY OF CRAIOVA, “EUGENIU CARADA” DOCTORAL SCHOOL OF ECONOMIC SCIENCES, CRAIOVA, ROMANIA
Stefan Margaritescu: DOCTORAL SCHOOL OF ECONOMIC SCIENCES ”EUGENIU CARADA”, UNIVERSITY OF CRAIOVA, CRAIOVA, ROMANIA

Annals - Economy Series, 2026, vol. 1, 7-25

Abstract: The accurate prediction of financial time-series remains a formidable challenge due to inherent non-linearity, heteroskedasticity, and the presence of "fat-tailed" distributions. This study proposes a novel, risk-augmented hybrid ensemble framework designed to enhance the forecasting precision of the Amsterdam Exchange (AEX) index. Departing from conventional monolithic models, the research methodology integrates an asymmetric power autoregressive conditional heteroskedasticity (APARCH) model with a Student-t distribution to extract robust volatility features. These econometric inputs are subsequently fed into a tripartite deep learning ensemble comprising Long Short-Term Memory (LSTM) networks for temporal dependencies, Convolutional Neural Networks (CNN) for spatial feature extraction, and Random Forest (RF) for non-linear regression refinement. Empirical results demonstrate that the proposed architecture significantly outperforms baseline models, achieving a high predictive accuracy characterized by an $R^{2}$ of 0.9408 and a Mean Absolute Error (MAE) of 4.9542. A critical finding of this research is the significance of the leptokurtic nature of AEX returns (Kurtosis: 12.64); by anchoring the machine learning engine with APARCH-derived conditional volatility, the model effectively mitigates the impact of market noise and transient shocks. Furthermore, Value-at-Risk (VaR) backtesting validates the model’s reliability for risk management, revealing that actual violations (181) remained well below the theoretical expectations (249.4) at a 95% confidence interval. The study concludes with a 30-day forward volatility projection, offering actionable insights for institutional investors and policy-makers during periods of market transition. By bridging the gap between classical econometrics and advanced computational intelligence, this research provides a robust template for multi-stage financial forecasting in volatile global markets.

Keywords: AEX Index Forecasting; Hybrid Ensemble Learning; LSTM-CNN Architecture; Random Forest; APARCH-t Volatility; Financial Time-Series Analysis; Leptokurtic Returns; Risk-Aware Deep Learning; Value-at-Risk (VaR); Conditional Volatility. (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.utgjiu.ro/revista/ec/pdf/2026-01/01_Aman.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:cbu:jrnlec:y:2026:v:1:p:7-25

Access Statistics for this article

More articles in Annals - Economy Series from Constantin Brancusi University, Faculty of Economics Contact information at EDIRC.
Bibliographic data for series maintained by Ecobici Nicolae ().