基于关系增强图的社交机器人检测方法、装置、设备及存储介质
By constructing a multi-relationship enhanced social network and processing user data from social platforms using a multi-module neural network, topic-emotion binary pairs and semantic behavior fingerprints are generated, solving the problem of difficulty in identifying covert social bots in existing technologies and achieving high-precision social bot detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 湖南工商大学
- Filing Date
- 2026-03-31
- Publication Date
- 2026-07-17
AI Technical Summary
Existing social bot detection technologies struggle to identify covert social bots that possess semantic camouflage capabilities and lack direct interaction relationships. Traditional methods fail to capture deep semantic connections in tweets or ignore implicit semantic relationships, resulting in low detection accuracy.
By acquiring the basic account attributes, account interaction relationships, and historical tweet content of social platform users, a multi-agent collaborative mechanism is constructed using a large language model to parse tweet content, generate topic-sentiment binary pairs, construct a multi-relationship enhanced social network, calculate similarity by combining semantic behavior fingerprints, process account attributes and semantic behavior fingerprints using a multi-module neural network, reconstruct feature weights and perform adaptive boundary determination, and output social robot detection results.
It significantly improves the detection accuracy of covert social robots, and can identify social robots with semantic camouflage capabilities but lacking direct interaction relationships, thus reducing the false alarm rate.
Smart Images
Figure CN121935735B_ABST