Recurrent convolutional neural kernel model for stock price movement prediction.

Area of Science:

• Artificial Intelligence
• Computational Finance
• Machine Learning

Background:

• Stock price movement prediction is crucial for investment decisions.
• Traditional methods often treat it as a binary classification task.
• Integrating diverse data sources can enhance prediction accuracy.

Purpose of the Study:

• To propose a novel Recurrent Convolutional Neural Kernel (RCNK) model for stock price prediction.
• To leverage both historical price data and sequential text sentiment data.
• To improve prediction accuracy by combining technical and sentiment analysis.

Main Methods:

• Developed a Recurrent Convolutional Neural Kernel (RCNK) model.
• Treated text data as sequential input for sentiment embeddings with temporal features.
• Employed an explicit kernel mapping layer to reduce model parameters and overfitting.
• Compared RCNK against baseline deep learning models and single-data-source models.

Main Results:

• The proposed RCNK model demonstrated superior performance in stock price movement prediction.
• Integrating historical prices with sequential sentiment analysis yielded significant improvements.
• The explicit kernel mapping layer effectively reduced model complexity and overfitting risks.

Conclusions:

• The RCNK model offers a powerful approach for stock price prediction by synergizing technical and sentiment analysis.
• Treating sentiment data sequentially and using explicit kernel mapping enhances model effectiveness.
• This integrated approach provides a more robust and accurate method for financial market forecasting.