Paper: arXiv 2404.01624

Authors: Qishuo Cheng

Abstract

In recent decades, financial quantification has emerged and matured rapidly. For financial institutions such as funds, investment institutions are increasingly dissatisfied with the situation of passively constructing investment portfolios with average market returns, and are paying more and more attention to active quantitative strategy investment portfolios. This requires the introduction of active stock investment fund management models. Currently, in my country’s stock fund investment market, there are many active quantitative investment strategies, and the algorithms used vary widely, such as SVM, random forest, RNN recurrent memory network, etc. This article focuses on this trend, using the emerging LSTM-GRU gate-controlled long short-term memory network model in the field of financial stock investment as a basis to build a set of active investment stock strategies, and combining it with SVM, which has been widely used in the field of quantitative stock investment. Comparing models such as RNN, theoretically speaking, compared to SVM that simply relies on kernel functions for high-order mapping and classification of data, neural network algorithms such as RNN and LSTM-GRU have better principles and are more suitable for processing financial stock data. Then, through multiple By comparison, it was finally found that the LSTM- GRU gate-controlled long short-term memory network has a better accuracy. By selecting the LSTM-GRU algorithm to construct a trading strategy based on the Shanghai and Shenzhen 300 Index constituent stocks, the parameters were adjusted and the neural layer connection was adjusted. Finally, It has significantly outperformed the benchmark index CSI 300 over the long term. The conclusion of this article is that the research results can provide certain quantitative strategy references for financial institutions to construct active stock investment portfolios.

Complexity vs Empirical Score

  • Math Complexity: 3.5/10
  • Empirical Rigor: 2.0/10
  • Quadrant: Philosophers — conceptual discussion, lower math and empirical scores

Why this score: The paper discusses deep learning architectures (LSTM-GRU, RNN) and mentions mathematical concepts like gradient issues and activation functions, but lacks heavy derivations or dense formulas. Empirically, it only describes a backtest on CSI 300 with vague details on implementation, data, or robust metrics, making it not backtest-ready.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Optimize Active Stock<br>Quantitative Strategies"] --> B["Data Input:<br>Shanghai & Shenzhen<br>300 Index Stocks"]
  B --> C["Methodology Comparison:<br>SVM vs RNN vs<br>LSTM-GRU Models"]
  C --> D{"Computational Process:<br>Model Training &<br>Parameter Tuning"}
  D --> E["Key Finding 1:<br>LSTM-GRU Achieves<br>Highest Accuracy"]
  D --> F["Key Finding 2:<br>Strategy Outperforms<br>CSI 300 Benchmark"]
  E --> G["Outcome:<br>Intelligent Optimization<br>Framework for<br>Financial Institutions"]
  F --> G

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a theoretical framework and trading strategy based on an LSTM‑GRU deep learning model for active stock investment.

Source sentence
Finally, we proposed a theoretical framework for stock trading strategy design based on the LSTM‑GRU deep learning model and conducted experimental verification.

Assumes

No explicit assumptions found in the text.

Data

  • CSI 300 Index market-prices unknown
    CSI 300 constituents
    Source sentence
    Through empirical research and experimental verification on the A‑share CSI 300 Index, we have drawn the following conclusions
  • Shanghai and Shenzhen 300 Index constituent stocks market-prices unknown
    Shanghai and Shenzhen 300 Index constituents
    Source sentence
    By selecting the LSTM‑GRU algorithm to construct a trading strategy based on the Shanghai and Shenzhen 300 Index constituent stocks

Costs

Not modeled A trading or execution result is reported without costs.

Finds 3

  • The LSTM‑GRU model significantly outperforms the benchmark CSI 300 index over the long term. performance
    Source sentence
    Finally, It has significantly outperformed the benchmark index CSI 300 over the long term.
  • LSTM models show better modeling effect and prediction accuracy compared to traditional SVM. predictive-accuracy
    Source sentence
    Compared with traditional machine learning algorithms such as SVM, the LSTM model shows better performance in both modeling effect and prediction accuracy.
  • The proposed LSTM‑GRU framework outperforms passive index stock funds with better risk control and stability. performance
    Source sentence
    Experimental results show that this framework can significantly outperform passive index stock funds in the long term, while having better risk control capabilities and stability.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 29 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (6)

  1. ImageNet classification with deep convolutional neural networks (2017) — link
  2. A Novel Bearing Fault Classification Method Based on XGBoost: The Fusion of Deep Learning-Based Features and Empirical Features (2020) — link
  3. Adversarially learned one-class novelty detection with confidence estimation (2020) — link
  4. Error Analysis and Accuracy Calibration Method of U-Tube Coriolis Mass Flowmeter Under Pulsating Flow (2021) — link
  5. IEEE Transactions on Image Processing (2023) — link
  6. Ecological water security impact of large coal base development and its protection (2023) — link

0 of 6 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.