write in front
AAAI 2023 will be held in Washington, DC, USA from February 7 to February 14, 2023. This conference received a total of 8777 submissions and accepted 1721 papers, with an acceptance rate of 19.6%. This article introduces several quantitative trading-related papers included in AAAI 2023.
Paper title:
Financial Time Series Forecasting using CNN and Transformer
Author's Affiliation:
JP Morgan AI research
Paper link:
https://arxiv.org/pdf/2304.04912.pdf
research content:
Time series forecasting is important for decision making in various fields. And financial time series like stock prices are difficult to predict because it is difficult to model short- and long-term time dependencies between data points . Convolutional neural networks (CNN) are good at capturing local patterns and are used to model short-term dependencies. However, due to the limited acceptance range, CNN cannot learn long-term dependencies. Transformer, on the other hand, is able to learn global context and long-term dependencies. In this article, the author proposes to take advantage of CNN and Transformer to simultaneously simulate short-term and long-term dependencies in time series and predict whether future prices will increase, decrease, or remain unchanged (stationary). In experiments, the authors demonstrate the success of the proposed method in predicting intraday stock price movements of S&P 500 constituent stocks compared with commonly used statistical and deep learning methods.
model framework
Experimental results (trend classification)
Paper title:
PEN: Prediction-Explanation Network to Forecast Stock Price Movement with Better Explainability
Author's Affiliation:
Hillhouse School of Artificial Intelligence, Renmin University of China
Paper link:
https://ojs.aaai.org/index.php/AAAI/article/view/25648/25420
research content:
Nowadays, among banks, hedge funds and asset managers, the interpretability of stock price movement forecasts is receiving increasing attention for audit or regulatory reasons. Textual data, such as financial news and social media posts, may be partly responsible for stock price movements. To this end, the authors propose a novel prediction-explanation network (PEN) framework that jointly models and aligns text flow and price flow. The key component of the PEN model is a shared representation learning module that learns which texts may be relevant to stock price changes by simulating the interaction between text data and stock price data , using a salient vector to describe their correlation. In this way, the PEN model is able to predict stock price movements by identifying and utilizing rich information, while on the other hand, the selected text information also explains stock price movements. Experiments on real-world data sets show that it can kill two birds with one stone: in terms of accuracy, the proposed PEN model outperforms the baseline model; in terms of interpretability, the PEN model is proven to be far better than the attention mechanism and can very confidently Pick out key texts.
model framework
Experimental results (classification accuracy, Matthews correlation coefficient)
Paper title:
Optimal Execution via Multi-Objective Multi-Armed Bandits
Author's Affiliation:
University College London
Paper link:
https://ojs.aaai.org/index.php/AAAI/article/download/26945/26717
research content:
When attempting to liquidate a large position in a particular stock, the price of that stock is likely to be affected by the trade, resulting in the expected return being likely to be lower if we sold the entire position at once. This raises the question of optimized execution, which aims to split a sales order into several smaller sales orders over a period of time in order to achieve an optimal balance between share price and market risk. This problem can be defined in terms of difference equations. Here, the authors show how to reframe it as a multi-objective problem and then use a novel Multi-Armed Bandit Algorithm to solve it.
Algorithm process
Paper title:
Sequential Graph Attention Learning for Predicting Dynamic Stock Trends
Author's Affiliation:
Taipei National University
Paper link:
https://arxiv.org/pdf/2301.10153
research content:
The stock market is characterized by complex relationships between companies and the market. This research combines a sequence graph structure with an attention mechanism to learn global and local information within time series. Specifically, the “GAT-AGNN” module proposed by the authors compares model performance across multiple industries as well as within a single industry. The results show that on the Taiwan stock dataset, the proposed framework outperforms multiple benchmark methods in predicting stock trends across multiple industries.
model framework
Experimental results (classification and regression performance)
Recommended reading in previous issues
WWW 2023 | Papers related to quantitative trading (with paper links)
KDD 2023 | Papers related to quantitative trading (with paper links)
AAAI 2022 | Papers related to quantitative trading (with paper links)
IJCAI 2022 | Papers related to quantitative trading (with paper links)
WWW 2022 | Papers related to quantitative trading (with paper links)
KDD 2022 | Papers related to quantitative trading (with paper links)
Interpretation: Application of ChatGPT in stock market prediction
Interpretation: What kind of metrics should be used for machine learning prediction revenue models?
[Python Quantification] Deep learning model quantification backtesting framework based on backtrader
[Python Quantification] Use Transformer model for stock price prediction
[Python Quantification] Build a CNN-LSTM model for stock price prediction
[Python Quantification] Use Python to build a stock public opinion analysis system
[Python Quantification] Use Informer for stock price prediction
[python quantification] Using DeepAR for multi-step probability prediction of stock prices
"Artificial Intelligence Quantitative Laboratory" Knowledge Planet
By joining the Knowledge Planet of the Artificial Intelligence Quantitative Laboratory, you can get: (1) Regular push of the latest research results related to the quantitative application of artificial intelligence, including high-level journal papers and high-quality financial engineering research reports from securities companies, so that you can understand the latest cutting-edge knowledge anytime and anywhere; (2) The complete source code of the Python project in the official account’s historical articles; (3) High-quality Python, machine learning, and quantitative trading related e-books PDF; (4) High-quality quantitative trading information and project code sharing; (5) Communicate and make friends with star friends Like-minded friends. (6) Ask the bloggers questions and answer questions.
