STORM: A Spatio-Temporal Factor Model Based on Dual Vector Quantized Variational Autoencoders for Financial Trading

TL;DR

STORM introduces a spatio-temporal latent factor model based on dual VQ-VAE to learn cross-sectional and time-series stock factors as multi-dimensional embeddings. Through discrete codebooks, cross-attention-based fusion, and prior-posterior learning, STORM achieves orthogonal, diverse factors that improve return prediction and trading performance. Empirical results on SP500 and DJ30 show STORM outperforms baselines in portfolio management and algorithmic trading, while delivering stronger factor quality (RankIC/RankICIR) and robust risk-adjusted returns. The approach is scalable and practical for real-world trading, with plans to incorporate exogenous information such as news in future work.

Abstract

In financial trading, factor models are widely used to price assets and capture excess returns from mispricing. Recently, we have witnessed the rise of variational autoencoder-based latent factor models, which learn latent factors self-adaptively. While these models focus on modeling overall market conditions, they often fail to effectively capture the temporal patterns of individual stocks. Additionally, representing multiple factors as single values simplifies the model but limits its ability to capture complex relationships and dependencies. As a result, the learned factors are of low quality and lack diversity, reducing their effectiveness and robustness across different trading periods. To address these issues, we propose a Spatio-Temporal factOR Model based on dual vector quantized variational autoencoders, named STORM, which extracts features of stocks from temporal and spatial perspectives, then fuses and aligns these features at the fine-grained and semantic level, and represents the factors as multi-dimensional embeddings. The discrete codebooks cluster similar factor embeddings, ensuring orthogonality and diversity, which helps distinguish between different factors and enables factor selection in financial trading. To show the performance of the proposed factor model, we apply it to two downstream experiments: portfolio management on two stock datasets and individual trading tasks on six specific stocks. The extensive experiments demonstrate STORM's flexibility in adapting to downstream tasks and superior performance over baseline models.

Figures (3)

• Figure 1: Overall architecture of STORM
• Figure 2: The architecture of our proposed STORM model
• Figure 3: (a) Distribution of counts for codebook embedding indices. (b) Hyperparameter experiment results for different codebook sizes.