Dynamic Financial Statement Anomaly Detection System

Develop an advanced anomaly detection system using Python that automatically scans and flags potential financial reporting irregularities across multiple corporations. Implement unsupervised machine learning techniques like Isolation Forest and Local Outlier Factor to identify statistically significant deviations in balance sheets, income statements, and cash flow statements. The system should generate detailed investigative reports with probability-weighted anomaly scores and recommended follow-up actions.