Research On Evaluating System Of Express Enterprise Based On Data Mining Technology And Analytic Hierarchy Process

The rapid development of the logistics and express industry has enabled a large number of express delivery companies to emerge constantly,providing consumers with services and products that are not well-balanced.In order to comply with consumers' basic expectations and demands for high-quality and low-cost express delivery services and good user experience,the evaluation and management of express delivery companies based on consumers' wishes is imperative.This paper uses data mining technology to design a set of architecture for evaluating express delivery companies.Based on the evaluation system,the analytic hierarchy process is used to help consumers make decisions before courier consumption.Referred to SERQUAL,LSQ and other evaluation models,combined with China's "Express Servicet" industry standards and relevant scholars'views,and based on consumer comments,this paper constructed a logistics express evaluation index system suitable for the actual situation.Taking the national post office complaint website as the data source platform,taking four express delivery companies,EMS,SF,YT,YS as examples,we obtained 4228 cases of complaints and praises from consumers for express delivery enterprises,and realized establishment of the evaluation system for express delivery enterprises based on data mining technology and analytic hierarchy analysis.The implementation process mainly includes the following parts.Text Categorization.We tried the four commonly used classification algorithms such as SVM,NNET,TREE,RM,comprehensively compared the accuracy,recall rate and F-score of the experimental results,then select the random forest algorithm with better effect as the text classification algorithm.The consumer comment text is divided into a certain dimension in the evaluation index system it reflects.Emotion Analysis.We have collected and constructed sentiment dictionaries such as positive/negative emotions and evaluation words,adverbs/negative words,and summarizes the emotion expression formula according to Chinese grammar expression habits.Based on the classic dictionary matching mode,we obtained the consumer evaluation sentiment scores by different enterprises under different dimensions of the evaluation index system.Decision Model.The traditional analytic hierarchy process relies on expert scoring to form an evaluation matrix,which has certain subjectivity and one-sidedness.This paper improved the evaluation form of AHP,transformed the actual needs and preferences of consumers into the evaluation matrix of the program layer and the criteria layer,converted the sentiment score of the express delivery company into the evaluation matrix of the criterion layer and the plan layer,and then weights of each program are calculated to make the optimal decision.