Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MambaStock: Selective state space model for stock prediction

Zhuangwei Shi

TL;DR

The paper tackles nonlinear volatility in stock market time series and limitations of ARIMA. It introduces MambaStock, a stock predictor built on the Mamba framework with a selective state-space architecture (S6). Across four Chinese stocks using Tushare data, MambaStock outperforms a wide set of baselines including KF, ARIMA, ARIMA-NN, XGBoost, LSTM variants, Transformer, TL-KF, and AttCLX, as measured by standard metrics. The work demonstrates the practical potential of Mamba for finance time-series forecasting and provides reproducible resources through public data and accessible source code.

Abstract

The stock market plays a pivotal role in economic development, yet its intricate volatility poses challenges for investors. Consequently, research and accurate predictions of stock price movements are crucial for mitigating risks. Traditional time series models fall short in capturing nonlinearity, leading to unsatisfactory stock predictions. This limitation has spurred the widespread adoption of neural networks for stock prediction, owing to their robust nonlinear generalization capabilities. Recently, Mamba, a structured state space sequence model with a selection mechanism and scan module (S6), has emerged as a powerful tool in sequence modeling tasks. Leveraging this framework, this paper proposes a novel Mamba-based model for stock price prediction, named MambaStock. The proposed MambaStock model effectively mines historical stock market data to predict future stock prices without handcrafted features or extensive preprocessing procedures. Empirical studies on several stocks indicate that the MambaStock model outperforms previous methods, delivering highly accurate predictions. This enhanced accuracy can assist investors and institutions in making informed decisions, aiming to maximize returns while minimizing risks. This work underscores the value of Mamba in time-series forecasting. Source code is available at https://github.com/zshicode/MambaStock.

MambaStock: Selective state space model for stock prediction

TL;DR

The paper tackles nonlinear volatility in stock market time series and limitations of ARIMA. It introduces MambaStock, a stock predictor built on the Mamba framework with a selective state-space architecture (S6). Across four Chinese stocks using Tushare data, MambaStock outperforms a wide set of baselines including KF, ARIMA, ARIMA-NN, XGBoost, LSTM variants, Transformer, TL-KF, and AttCLX, as measured by standard metrics. The work demonstrates the practical potential of Mamba for finance time-series forecasting and provides reproducible resources through public data and accessible source code.

Abstract

The stock market plays a pivotal role in economic development, yet its intricate volatility poses challenges for investors. Consequently, research and accurate predictions of stock price movements are crucial for mitigating risks. Traditional time series models fall short in capturing nonlinearity, leading to unsatisfactory stock predictions. This limitation has spurred the widespread adoption of neural networks for stock prediction, owing to their robust nonlinear generalization capabilities. Recently, Mamba, a structured state space sequence model with a selection mechanism and scan module (S6), has emerged as a powerful tool in sequence modeling tasks. Leveraging this framework, this paper proposes a novel Mamba-based model for stock price prediction, named MambaStock. The proposed MambaStock model effectively mines historical stock market data to predict future stock prices without handcrafted features or extensive preprocessing procedures. Empirical studies on several stocks indicate that the MambaStock model outperforms previous methods, delivering highly accurate predictions. This enhanced accuracy can assist investors and institutions in making informed decisions, aiming to maximize returns while minimizing risks. This work underscores the value of Mamba in time-series forecasting. Source code is available at https://github.com/zshicode/MambaStock.
Paper Structure (9 sections, 11 equations, 5 figures, 4 tables)

This paper contains 9 sections, 11 equations, 5 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Structured state space sequence model (S4)
  4. Mamba
  5. MambaStock
  6. Experiments
  7. Prediction performance
  8. Compared with other methods
  9. Conclusions

Figures (5)

  • Figure 1: MambaStock framework.
  • Figure 2: Prediction on 600036.SH
  • Figure 3: Prediction on 601288.SH
  • Figure 4: Prediction on 601328.SH
  • Figure 5: Prediction on 601988.SH