Research And Application Of Knowledge Mapping For Public Safety

With the continuous deepening of knowledge graph technology research in recent years,all walks of life have gradually introduced this new type of data processing technology.Knowledge maps provide more valuable decision support for the field.In public safety,the knowledge map can stitch together the various features of the tracking object.Through the map,various scattered information is linked together to form a complete information network,helping staff to quickly locate problems and find solutions.Therefore,the systematic research in this paper is based on the knowledge map construction in the security field,and the relationship extraction method is improved.This paper analyzes in detail the process and method of knowledge graph construction.And this paper designed a new relationship extraction algorithm to improve the accuracy of entity relationship extraction.Knowledge graphs are mostly applied in a vertical field.However,the existing relationship extraction methods only focus on the extraction algorithm itself,but ignore the domain-specific expressions to help quickly extract the entity relationship corresponding to the sentence.The semi-supervised learning method only needs a small amount of labeled corpus to complete more accurate extraction results.Therefore,this paper proposes an improved method of using domain expression features and Boot Strapping algorithm fusion.Extract the corresponding special simple sentence patterns in the field from the case description text set.Then according to the syntax analysis,the entity relationship structure corresponding to the sentence pattern is obtained.If a sentence in the test data set conforms to the extracted domain specific structure,the extracted sentence pattern is used to extract the entity relationship.Otherwise,it is processed and added to the pending set to use the Boot Strapping algorithm for entity relationship extraction.In order to verify the effectiveness of the method,the experimental text will be obtained from the network for cleaning.At the same time,some corpora are extracted for manual labeling,which is convenient for later verification algorithm results.The experimental results show that the improved algorithm has a better improvement than the simple Boot Strapping algorithm.Finally,this paper analyzes the application of case search platform based on knowledge map.