VertiFuseX improves stock prediction by combining multiple time methods
VertiFuseX: Generalizable Financial Forecasting via Multi-Stream Temporal Fusion
Machine Learning
Summary
Predicting stock prices is hard because financial data changes a lot and has noise. The authors created VertiFuseX, a model that mixes different timelines of stock data at an intermediate step instead of the usual final step to keep more useful information. They tested it on 15 years of global stock data and found it predicts prices more accurately and is better across different markets than other recent models. The model also helps reduce financial risks when tested with simulated trading during extreme market conditions.
What this means in practice
- •For financial analysts: Improve stock market forecasts by integrating multi-scale temporal features for more accurate and robust predictions under various market conditions.
- •For algorithmic trading teams: Develop trading strategies that reduce maximum losses and improve returns by using a fusion-based forecasting model tested under extreme market regimes.
Authors
Aashish Bohra, Vivek Vijay
Abstract
Stock price prediction remains challenging due to the non-stationary and noisy nature of financial time series. Existing deep learning models often rely on rigid decision-level fusion, ad hoc hyperparameter tuning, and compressed final-layer outputs, causing information loss, overfitting, and limited cross-market generalization. We propose VertiFuseX, a hybrid LSTM architecture using penultimate-layer vertical fusion of multi-scale temporal representations. VertiFuseX stacks and reweights penultimate features from LSTM, Bi-LSTM, and St-LSTM branches, integrates a parallel DNN stream, and jointly optimizes all components via backpropagation under a fixed hyperparameter configuration. This preserves richer intermediate temporal information across scales. Evaluated on 15 years (2010-2024) of closing prices from 10 global equity indices using strict chronological out-of-sample testing with the final 365 trading days held out, VertiFuseX achieves 30-54% MAPE reductions and over 40% improvements in MAE and RMSE versus LSTM-based baselines, and outperforms seven state-of-the-art models across 33 metric-dataset comparisons. Ablation studies confirm penultimate-layer fusion drives these gains over final-layer fusion and decision-level ensembling. Gradient-based saliency analysis shows consistent emphasis on mid-range dependencies at lags 9-15 days. Economic validation via algorithmic trading simulation under extreme market regimes shows reduced maximum drawdowns and superior risk-adjusted returns. With 675k parameters, a 2.6 MB memory footprint, and 1.5 ms/sample inference latency, VertiFuseX offers a lightweight, interpretable, deployment-ready framework for robust financial forecasting.