A novel system for sensitive topic detection and alert assessment

This paper proposes a novel system for sensitive topic detection and alert assessment which is based on dynamic knowledge base (DKB) and cooperative methods. Multi-node cooperative topic detection model (MCTDM) can not only detect new sensitive topic within single domain, but also conduct information cooperative scheduling among several DKBs, thus it has strong ability to detect new sensitive topic and burst sensitive topic. Alert assessment based on cooperative computing (AACC) evaluates the impact of alert topic on both internal and external domain. Through cooperatively scheduling high frequency alert topic case among different DKBs, it effectively improves the ability of alert topic spreading analysis and shortens the time span of alert topic spreading. Therefore, DKB can dynamically update and gradually enrich. The system combines MCTDM, AACC and DKB, which already runs in several colleges and universities. During the period of system application, it forecasts several alert topics in time and achieves a good application result which verifies the effectiveness and availability of above methods.