EEG-Based Emotion Classification in Financial Trading Using Deep Learning: Effects of Risk Control Measures

Area of Science:

• Neuroscience and Computational Finance
• Application of deep learning in financial market analysis
• EEG-based emotion recognition

Background:

• Day traders face high pressure, influencing decisions and leading to losses.
• Emotional states significantly affect trading outcomes, especially in volatile markets.
• The impact of risk control measures on trader emotions remains unclear.

Purpose of the Study:

• To assess the impact of stop loss and limit orders on day trader emotions.
• To develop a deep learning framework for emotion classification in financial trading using EEG data.
• To compare emotional states during trading with and without risk control measures.

Main Methods:

• Two experiments were conducted: one with stop loss/limit orders, one without.
• A novel hybrid neural network (CNN-BiLSTM-CRF) with Bayesian Optimization was developed.
• Electroencephalography (EEG) data was used for valence-arousal emotion classification.

Main Results:

• The proposed model achieved high classification accuracies (85.65% and 85.05%).
• Trading with stop loss/limit orders correlated with High Valence/High Arousal emotions (hope).
• Trading without these measures increased Low Valence/High Arousal emotions (fear, worry).
• Calmness (High Valence/Low Arousal) was highest in a non-trading control group.

Conclusions:

• The developed framework effectively classifies emotions in financial trading.
• Stop loss and limit orders positively influence trader emotional states.
• Findings can enhance risk management and decision-making for day traders.