Outline a methodology for detecting anomalies in real-time high-frequency trading data. Compare and contrast the efficacy of Isolation Forests, Autoencoders, and Statistical Process Control for this specific use case. Discuss how you would handle the class imbalance problem inherent in fraud detection and propose a validation metric that accounts for the cost of false positives versus false negatives.