| The outbreak of the COVID-19 in December 2019 has had a tremendous impact on China's economy and the lives of residents.During the epidemic,people were restricted to go out.Therefore,the offline demand that was unable to be transferred to the online has caused great changes in the volume of the express business in China.In the context of the COVID-19,this article analyzes the temporal and spatial characteristics of the distribution of the number of people diagnosed with the COVID-19 in 31 provinces in mainland my country,as well as the difference in express delivery business volume in my country before and after the COVID-19.On this basis,it is considered to add the influencing factors of the COVID-19 to the analysis of the impact on the express business volume of mainland my country,which makes up for the gap in this research field,and at the same time from the regional,economic level,consumption level and residents' income Other characteristic variables were selected from the four types of influencing factors.From the perspective of statistical analysis,a regression analysis model with express delivery business volume as the dependent variable was established,and the relationship between express delivery business volume and each characteristic variable was determined,and the various characteristics were analyzed.The mode and degree of the influence of characteristic variables on the express business volume of each province.At the same time,this paper takes machine learning methods as another entry point,and establishes a courier business volume fitting model based on Decision Tree,Support Vector Machine Regression(SVR),Random Forest,GBDT,and analyzes the COVID-19 and others from the perspective of machine learning.The impact of influencing factors on the express business volume,the idea of the machine learning model is to divide the collected data into a training set and a test set to be used for model training and fitting effect testing,through the adjustment of each parameter in each model Finally,a model with a good fitting effect was obtained.Finally,the prediction accuracy of each model was compared with the mean square error under the same test set,and the fitting effect of each model was analyzed.In order to further improve the fitting effect of the model,this article introduces the stacking integrated learning method.The previous five models are used as the base learner,and the data processed by the base learner is trained again to obtain the optimal express business volume fitting model Finally,the optimality of the integrated model was verified by comparing the mean square error of the test set,and the final result of the model was comprehensively analyzed.Due to the certain difference in the domestic outbreak time and location of the COVID-19,it is also integrated Taking into account the influence of factors such as economic development level and residents' income,the final fitting model is not applicable to all provinces,and different models are applicable to different provinces.The final conclusion of this article is for domestic express delivery.The development of the industry puts forward several suggestions to provide certain reference and reference significance for local governments and express companies in China in resource arrangements and business development decisions in the post-epidemic period and in emergencies that may be faced in the future.This paper contains 25 figures,9 tables,and 60 references. |
PDF Full Text Request