Hybrid Autoencoder–LSTM Deep Learning Model for Detecting Anomalous Behavior in Metaverse Financial Transactions
Journal of Digital Market and Digital Currency · 02.11.2026 · 3 (4) · 299-314
DOI: 10.47738/jdmdc.v3i4.103 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Digital Market and Digital Currency, online erschienen am 02.11.2026. Verfasst von Amaleswari Rajulapati.
Abstract (Originalsprache)
The expansion of the metaverse has created new possibilities for digital financial interaction while simultaneously increasing exposure to fraudulent and anomalous activities. This study introduces a Hybrid Autoencoder–LSTM Deep Learning Model designed to detect irregular transaction behaviors within metaverse financial systems. The model integrates an Autoencoder for nonlinear feature extraction and an LSTM network for sequential dependency learning, enabling comprehensive detection of behavioral and transactional anomalies. Using a dataset of 78,600 records encompassing demographic, behavioral, and financial attributes, the model achieved an overall accuracy of 79.76 percent and an Area Under the Curve (AUC) value of 0.742, demonstrating moderate capability in distinguishing legitimate from fraudulent transactions. The convergence of training and validation losses confirmed robust generalization without overfitting, while classification metrics revealed strong precision and recall for legitimate transactions but reduced sensitivity to rare anomalies due to class imbalance. These results validate the effectiveness of the proposed hybrid model as a foundation for intelligent fraud detection in virtual financial environments and highlight the importance of integrating deep learning and explainable AI approaches to strengthen security, transparency, and trust within metaverse economies.
Zitieren
APA
Rajulapati, A. (2026). Hybrid Autoencoder–LSTM Deep Learning Model for Detecting Anomalous Behavior in Metaverse Financial Transactions. Journal of Digital Market and Digital Currency, 3(4), 299-314. https://doi.org/10.47738/jdmdc.v3i4.103
Vancouver
Rajulapati A. Hybrid Autoencoder–LSTM Deep Learning Model for Detecting Anomalous Behavior in Metaverse Financial Transactions. Journal of Digital Market and Digital Currency. 2026;3(4):299-314. doi:10.47738/jdmdc.v3i4.103
BibTeX
@article{Rajulapati2026,
author = {Rajulapati, Amaleswari},
title = {Hybrid Autoencoder–LSTM Deep Learning Model for Detecting Anomalous Behavior in Metaverse Financial Transactions},
journal = {Journal of Digital Market and Digital Currency},
year = {2026},
volume = {3},
number = {4},
pages = {299--314},
doi = {10.47738/jdmdc.v3i4.103}
}
EUVIMED