Research And Application Of Person Relationship Extraction Method Based On Deep Learning

With the rapid development of Internet of Things technology,Internet users are creating huge amounts of network information all the time.Information extraction technology is used to quickly obtain the required content from massive information.It converts unstructured text data into structured text data,which is convenient for users to store and analyze.In the whole process of Information Extraction,Named Entity Recognition(NER)and Relationship Extraction(RE)are very important parts.In fact,Named Entity Recognition is the process of identifying the names of people,locations and organizations.Relationship extraction represents the relationship between entities in the form of triples.Due to the rapid development of deep learning,many scholars have started to adopt deep learning-based methods to deal with the problem of information extraction and achieved good results.Since neural networks have the capability of feature learning,so they can avoid the complex and time-consuming process of manually defining features in traditional methods.In this paper,two novel network models are proposed for the Chinese entity recognition and relationship extraction tasks respectively.These two novel network models improve the original models to enhance their model recognition effects,so that they can be better applied in Chinese information extraction.Bi Lattice-LSTM-Self ATT model is adopted in the task of Chinese Named Entity Recognition.The Chinese text is semantically encoded by using the Bi Lattice-LSTM method,so as to effectively solve the problem of word segmentation errors in the entity recognition process.Then the attention mechanism is utilized to weight the semantic coding,and finally the Conditional Random Field(CFR)is for decoding the semantics to complete the task of NER.The experimental results show that the Named Entity Recognition model proposed in this paper outperforms existing methods,with the method achieving an accuracy of 94.51%,a recall of93.13% and an F1 value of 93.81% on the dataset.The DP-BIGRU-ATT model is used in the task of Inter-personal Relation Extraction,which adds information like entity labeling and dependency relationship as external semantic features,thereby enhancing the semantic relationship information of the input text sequence.An attention mechanism is added to the model to focus on the areas that have a greater impact on the relationship extraction.The experimental results show that the Interpersonal Relationship Extraction model proposed in this paper outperforms existing methods,with the method achieving an accuracy of 0.61%,a recall of 63.54% and an F1 value of 74.70% on the dataset.