The Extracting Method For Classification Of School Labor Education Practice Based On Deep Learning

With the development of the information age,there are currently many labor education implementation cases on the Internet,and the implementation of labor education needs to be combined with the actual situation and adhere to the principle of adapting measures to local conditions.Labor education implementation cases are unstructured text data,including labor education participants and labor education event component categories and other information.This dissertation focuses on labor education events to describe the process of labor education development.This dissertation uses natural language processing methods to extract information such as schools,addresses,and Classification from labor education implementation cases,establishes a labor education Event-logic Graph,and analyzes the current situation and strategies of implementation according to local conditions.The main contents of this dissertation are:(1)A set of deep learning-based method framework for the extraction of school labor education practice components is proposed.There is a big difference between the research and knowledge extraction methods for the existing concept of the map of affairs and this dissertation,and it cannot be fully transferred to this dissertation.Therefore,it is necessary to construct a suitable method framework based on the characteristics of this research.In view of the lack of ontology construction schemes in the field of labor education and the lack of research on the value of unstructured labor education implementation case text information,this dissertation focuses on labor education events and builds a labor education Event-logic Graph ontology based on top-down ontology modeling ideas.Aiming at the lack of relevant corpus in the field of labor education to construct an Event-logic Graph,this dissertation designs a corresponding crawler solution to obtain the text data needed to build an Event-logic Graph from the Internet.(2)Screening of labor education case resources.The labor education cases shared on the Internet are important resources for the implementation and analysis of labor education according to local conditions,but the selection of resources faces challenges such as long report texts and low differentiation.this dissertation proposes a labor education case resource screening method based on two-terminal Attention convolutional neural network,which is used to extract local details and overall structural features related to cases from long texts.The experimental results show that the classification accuracy of the two-terminal Attention convolutional neural network for labor education case screening reaches 84.55%.(3)Type identification of labor education.This dissertation transforms the problem of labor education type identification into a multi-classification problem,it to confirm the category of labor education content.Due to the limitation of labor education development conditions,the proportions of various types of labor are inconsistent,resulting in class imbalance.This dissertation designs a network model of Attention mechanism combined with label text features,and reduces the impact of class imbalance by changing the embedded category feature vector.Compared with seven mainstream deep learning classification models,the results show that the model designed in this dissertation is superior to other models,and the classification accuracy rate reaches 82.82%.(4)Construction of labor education Event-logic Graph.Cases with specific implementation of labor education can be obtained through the screening of labor education case resources.This dissertation combines deep learning and natural language processing technology to extract information such as participants and participating locations,and uses the labor education type identification method to obtain the corresponding type.According to the extracted information,the Neo4 j graph database is used for storage and visualization,and the analysis of the Event-logic Graph is used as a reference for the development of labor education according to local conditions.