Based On LDA And FastText Under The COVID-19 Epidemic Logistics Theme Business And Matching Research

As an important part of my country's national economic development,the logistics industry has played an important role in promoting economic growth.The COVID-19(Corona Virus Disease 2019)is the most difficult public health emergency to prevent and control in recent years.Logistics companies,as transportation guarantee parties,play an important role in emergency management activities.However,logistics companies are facing unprecedented challenges during the COVID-19 epidemic.The country has successively introduced relevant logistics support policies,which to a certain extent alleviated the impact of the logistics industry during the epidemic.The country has successively introduced relevant logistics support policies,which to a certain extent alleviated the impact of the logistics industry during the epidemic.However,due to the suddenness and severity of the epidemic,the government may not be able to discover the demands of enterprises in a timely manner,resulting in a low matching rate between logistics policies and enterprise demands,resulting in poor policy implementation.Based on the above background,this article has launched a research on the matching of logistics policies and corporate demands under the COVID-19 epidemic:(1)Exploring the evolution trend of logistics policy themes during the epidemic period based on the LDA theme model.First,use the collected logistics policy texts during all stages of the epidemic period as training data to train the LDA(Latent Dirichlet Allocation)topic model;then,divide the policy texts during the epidemic period into stages,and input the divided logistics policy text into the trained LDA The document-topic distribution and topic-word distribution of each stage are obtained from the model,and the evolution trend of the theme of logistics policy during the epidemic period is explored;(2)Based on LDA and Fast Text,a phased study on the matching of logistics policies and corporate demands during the epidemic will be carried out.First,from the topic distribution of the logistics policy text obtained from the trained LDA model,the most probable one is selected as the core topic of the document,and the core topic assigned to each policy text is used as the text label;then,the labeled logistics policy is used as a training set to train the Fast Text neural network text classification model;finally,the trained Fast Text model is used to explore the core themes of corporate text.The study found that at various stages of the epidemic,the theme distribution of logistics support policies showed a certain degree of difference;the matching degree was poor in smooth government affairs,tax reduction and exemption,and transportation guarantee.Innovatively through the combination of LDA and Fast Text research methods can accurately and objectively judge the match between my country's logistics support policies and corporate demands during the epidemic,and provide reasonable policy recommendations for the scientific formulation of my country's logistics policies during the epidemic,thereby promoting Establish a scientific,reasonable and orderly logistics policy system.