Paper: arXiv 2412.01062
Authors: Yuxin Fan, Zhuohuan Hu, Lei Fu, Yu Cheng, Liyang Wang, Yuxiang Wang
Abstract
High-frequency trading (HFT) represents a pivotal and intensely competitive domain within the financial markets. The velocity and accuracy of data processing exert a direct influence on profitability, underscoring the significance of this field. The objective of this work is to optimise the real-time processing of data in high-frequency trading algorithms. The dynamic feature selection mechanism is responsible for monitoring and analysing market data in real time through clustering and feature weight analysis, with the objective of automatically selecting the most relevant features. This process employs an adaptive feature extraction method, which enables the system to respond and adjust its feature set in a timely manner when the data input changes, thus ensuring the efficient utilisation of data. The lightweight neural networks are designed in a modular fashion, comprising fast convolutional layers and pruning techniques that facilitate the expeditious completion of data processing and output prediction. In contrast to conventional deep learning models, the neural network architecture has been specifically designed to minimise the number of parameters and computational complexity, thereby markedly reducing the inference time. The experimental results demonstrate that the model is capable of maintaining consistent performance in the context of varying market conditions, thereby illustrating its advantages in terms of processing speed and revenue enhancement.
Complexity vs Empirical Score
- Math Complexity: 6.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper includes substantial mathematical derivations for feature selection and neural network pruning, indicating high math complexity, and it provides empirical results on real financial data (50 ETF) with implementation details like latency and performance metrics, showing high empirical rigor.
Research Flowchart
flowchart TD A["Research Goal: Optimize Real-Time Data Processing<br>in High-Frequency Trading Algorithms"] --> B["Data Input:<br>Real-Time Market Data"] B --> C["Methodology Step 1:<br>Dynamic Feature Selection"] C --> D["Computational Process:<br>Clustering & Feature Weight Analysis"] D --> E["Methodology Step 2:<br>Lightweight Neural Networks"] E --> F["Computational Process:<br>Fast Convolution & Pruning"] F --> G["Key Findings:<br>Reduced Inference Time &<br>Enhanced Revenue"]
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 model that optimizes real-time data processing in high‑frequency trading algorithms using dynamic feature selection and lightweight neural networks.
Source sentence
The objective of this work is to optimise the real-time processing of data in high-frequency trading algorithms.
Assumes 2
- The dynamic feature selection module uses k‑means clustering with a fixed number of clusters k=10. data
Source sentence
The dynamic feature selection module uses the k‑means clustering algorithm, the number of clusters k is set to 10, and the feature weights are adjusted once in each training cycle.
- The pruning threshold is set to 0.01 and the regularization parameter to 0.1, implying these values are appropriate for the model. model
Source sentence
In a lightweight neural network, the kernel sizes of the convolutional layer are 3×3 and 5×5, the pruning threshold is set to 0.01, and the regularization parameter is 0.1.
Data
- 50ETF dataset market-prices unknown
Source sentence
We selected 50 ETF datasets for model validation and evaluation. The 50ETF dataset contains the historical trading data of the 50 ETF on the Shanghai Stock Exchange, including the opening price, closing price, highest price, lowest price, trading volume and other characteristics.
Costs
Unclear
Finds 3
- The proposed method achieves the lowest RMSE across all parameter settings, indicating superior prediction accuracy and stability. performance
Source sentence
It can be seen that the Ours method exhibits the lowest RMSE across all parameters, indicating its superiority in prediction accuracy and stability.
- The method consistently attains high R² scores across a range of mutual information thresholds, showing stable accuracy. predictive-accuracy
Source sentence
As illustrated in Figure 3, the Ours method demonstrates consistent and high R² accuracy across a range of thresholds, exhibiting stability within the error range.
- The model maintains consistent performance under varying market conditions, enhancing processing speed and revenue. performance
Source sentence
The experimental results demonstrate that the model is capable of maintaining consistent performance in the context of varying market conditions, thereby illustrating its advantages in terms of processing speed and revenue enhancement.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 17 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.