Predictive modeling and anomaly detection in large-scale web portals through the CAWAL framework

2024-11-01

Knowledge-Based Systems

AI Summary

To address low accuracy and slow response in user behavior prediction and anomaly detection under dynamic topologies in large-scale web portals, this paper proposes CAWAL—a framework integrating temporal graph neural networks (T-GNNs) with adaptive online learning to jointly model long-range dependencies and topology evolution. It further introduces multi-granularity feature disentanglement and an unsupervised anomaly scoring mechanism, enabling end-to-end, label-free anomaly localization. Evaluated on a real-world million-scale log stream, CAWAL achieves 98.3% recall and 92.7% precision, with an average detection latency of only 3.2 seconds. These results demonstrate significant improvements in both the accuracy and timeliness of fault early warning for dynamic, large-scale web infrastructure.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Causality

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Algorithmic accountability and transparency on the web
• Web Mining and Content Analysis: Large pretrained models with web data

Problem

• User Behavior Prediction
• Anomaly Detection
• Website Management Efficiency

Innovation

• CAWAL framework
• Advanced Machine Learning Algorithms
• User Behavior Prediction