Research On The Classification And Quantification Of Policy Texts In The Field Of Digital Economy

Currently,the digital economy has become an important driving force for economic growth worldwide,and its strategic position continues to rise.Since the 18th National Congress of the Communist Party of China,various levels of government in China have successively issued a large number of policy documents supporting the high-quality development of the digital economy.However,due to the complexity and diversity of the digital economy,policy formulation and implementation face challenges.Therefore,research and analysis of digital economy policies are urgent tasks.With the continuous disclosure of government information,policy big data analysis provides a wealth of research materials.Meanwhile,the development of deep learning and natural language processing technologies provides technical support for intelligent interpretation of policy texts.This thesis aims to conduct text classification and quantification research on policy texts in the field of digital economy.Firstly,to address the problem of general word segmentation tools being unable to accurately and completely segment domain-specific vocabulary,a method of constructing a digital economy policy vocabulary was adopted,which included new word discovery algorithms,rule screening,and manual supplementation.Experimental results show that the policy vocabulary can accurately segment key information units in policy texts,with the accuracy,recall rate,and F1 value of text segmentation increasing by 24.49%,11.68%,and 19.69%,respectively.Secondly,from the perspective of decomposing digital economy policy tasks,this thesis constructed a text classification dataset and used three classic deep learning neural network models for automatic classification research,determining the best-performing deep learning model.Based on the policy vocabulary,the text segmentation and word vector training stages were improved,and a text classification method was proposed that integrates key features,significantly improving the evaluation indicators of text classification.By analyzing and classifying the development plan of digital economy for Shandong Province,this study reveals the weak links in the layout of Shandong’s digital economy.Additionally,in order to conduct more granular policy analysis,this thesis utilized the TF-IDF keyword algorithm for hotspot analysis of the national digital economy development plan and the Shandong digital strong province development plan.Furthermore,the BERTopic topic model was employed for topic modeling and visualization analysis of the digital economy policy text dataset,identifying 49 policy topics in different development task fields of the digital economy and comprehensively demonstrating China’s strategic layout for digital economy development.Finally,based on the research conclusions,this thesis proposes three policy implications to optimize the development of the digital economy in Shandong Province.Given the increasing volume of digital economy policy documents,this thesis leverages deep learning and NLP techniques to analyze digital economy policy texts,with the aim of enhancing information processing capabilities,reducing time and subjectivity in human interpretation,and promoting innovation in policy text big data processing.The thesis provides a reference for policy formulation and research.