BLSTM-CNN Relationship Classification Network Model

Relationship classification tasks are the process of extracting the relationship between two or more entities described in a natural language. Entity relationship classification can be used for search algorithms, structured knowledge extraction and knowledge map construction. The traditional relationship classification algorithm based on language characteristics can not meet the needs of entity relationship classification, and the use of deep learning method has become an important means to solve the task of relationship classification. In order to solve this problem effectively, the relationship classification model of Bi-LSTM-CNN structure is proposed. Experimental results show that the model has a high F1 score on the task data set of SemEval2010_task8 entity relationship extraction, which is better than the existing relationship classification task model.