Public crisis has the characteristics of suddenness and uncertainty, and it is necessary to combine the knowledge with the experience of other similar situations to make decisions effectively and quickly. This work combines artificial intelligent theory with information technology and brings case-based reasoning to build models consisting of the features of public crisis. We explore the case-representation approach and build a case-based retrieval algorithm. Combining the specificness of Case-based reasoning (CBR) technology in the monitoring of public crisis events, a new case retrieval algorithm for public crisis cases, named as Combined multi-similarity with set of simi-larity matching algorithm based on sememe (CMSBS), is proposed to analyze the cases with high similarity to current case. The CMSBS algorithm considers the structural and semantic similarities between two public crisis cases comprehensively. Simulation experiments are performed to validate the representation method of the knowledge, and the simulation results demonstrate that the CMSBS algorithm has superior performance in the average number of matching cases and matching accuracy rate and can work well in providing reference cases for subsequent events.
展开▼
