Deep Learning Architectures for Multi-Asset ETF Return Prediction and Portfolio Allocation: MLP, CNN-GAF, LSTM, GRU, and Ensemble Methods
“The GRU architecture delivers the highest overall directional accuracy (65.45% for SPY), outperforming LSTM on all metrics.”
AssetMulti-assetETFs
StrategyMachine learningmomentum
TechniqueRecurrent neural networkfeedforward network
Sample window2008–2022
Also stated: Sharpe ratio, Excess return, Drawdown, Long-short, Long-only, Price and volume, benchmark, out-of-sample test
Missing from abstract: costs, multiple testing, universe
Institutional-grade audit of this strategy.
£5 — Request the audit- Audit of the paper claims accounting for costs, luck, crises, originality, geography, etc.
- Reproduction code integratable into Bloomberg, IBKR, etc.
- Bias-free datasets for backtesting included
- Delivered in 2 working days
The paper's neighbourhoodMachine learning return prediction by year
The family is counted across the whole universe. Asset, method and authors are counted inside the family.