The proliferation of false flow in online platforms poses significant threats to user security and corporate governance due to its short lifespan and diverse fraudulent methods. To address the need for high reliability and real-time detection, this paper proposes an artificial intelligence (AI)-based active detection system. The system intelligently collects webpage URLs, calculates similarity using edit distance, and flags URLs exceeding a preset threshold as suspected targets for further analysis. Leveraging the ball vector machine (BVM) algorithm, the proposed method effectively classifies pre-processed feature vectors, achieving a detection accuracy of 95.6% and reducing the false alarm rate by 15% compared to support vector machine (SVM)-based methods. Experimental results on 500 legitimate and 500 fake webpages demonstrate its efficiency and suitability for large-scale applications, providing a robust and proactive solution to false flow detection.
