Research on Cross-border Securities Anomaly Detection Based on Time Zone Trading Characteristics

Cross-border securities trading presents unique challenges in financial market surveillance, particularly regarding temporal patterns that span multiple time zones. This research proposes a novel approach for detecting anomalous trading behaviors in international securities markets by leveraging time zone-specific trading characteristics. The methodology combines advanced machine learning techniques with temporal feature extraction to identify suspicious activities from Asia-Pacific region investors trading in US markets. The study analyzes trading patterns across Hong Kong, Singapore, and Australian time zones, developing a comprehensive framework that captures cultural and temporal nuances in trading behaviors. Experimental validation demonstrates significant improvements in anomaly detection accuracy while reducing false positive rates compared to traditional methods. The proposed system achieves 94.7% precision and 91.3% recall in identifying cross-border trading anomalies, with particular effectiveness in detecting after-hours suspicious activities. The research contributes to enhanced financial market integrity through culturally-aware AI models that support regulatory compliance across international trading platforms.