Predictive Modeling of Crypto Currency Trends Using Cloud-Deployed Ensemble Deep Learning

Read full text Summarise with VIRAT
Authors: Vishal Sresth, Amita Gandhi, Raghavendra Sai, Varanasi
Cited by: 0
References: 16

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Crypto-currency markets, particularly Bitcoin, exhibit significant price volatility due to the complex interplay of market demand, speculative trading behavior, and macroeconomic influences. While several recent studies have employed deep learning and attention-based models for Bitcoin trend prediction, few have integrated advanced feature extraction techniques with ensemble deep learning strategies in a scalable, cloud-based environment. Additionally, most prior works have focused on binary or regression-based outcomes, neglecting multi-class trend classification into uptrend, downtrend, and neutral categories. Addressing these gaps, the present study presents the cloud-deployed ensemble deep learning framework for crypto-currency trend prediction. This work incorporates a channel attention mechanism optimized by a improved watershed optimization (IWO) algorithm to improve the extraction of critical market features from both historical OHLCV (open, high, low, close, volume) data and computed technical indicators. These enriched features are then processed by an ensemble of deep learning models to enhance trend classification performance. Experimental results show that the proposed ensemble model significantly outperforms baseline models in terms of prediction accuracy, precision, recall, and F1-score for trend classification. The cloud deployment proved effective in delivering reliable and efficient processing for high-frequency financial data.

Fields of Study & Subjects

Stock Market Forecasting Methods Blockchain Technology Applications and Security FinTech Crowdfunding Digital Finance

Publication details

Publication date
9 February 2026
Language
EN
Document type
Preprint