Research And Implementation Of Data Analysis System For Express Complaint Business Based On Spark

With the development of the Internet and the continuous growth of China's GDP and foreign import and export volume,the express delivery industry has developed rapidly.In 2018,the national express delivery service business volume totaled 50.71 billion pieces,an increase of 26.6%year-on-year;however,due to various problems during the express delivery,users were delayed in receiving or dissatisfied with the process service.With the gradual improvement of the supervision of express delivery services,more and more users maintain their rights and interests through postal complaints,and the volume of complaints from the express delivery industry is on the rise.In view of the increasing volume of complaints in the postal industry,the low efficiency of manual processing and the promotion of real-name systems,this research analyzes large-scale data related to the complaint business and uses machine learning algorithms for data mining.Furthermore,to meet the needs of users,this research implements a data analysis system for express complaint business.This research will help improve the efficiency of the complaint business and provide reference for adjusting the regulatory policy,which has certain practical significance.The research points of this topic generally consist of four parts:1.Classification of complaint text:The complaint text is the text generated by the user on the complaint website describing the reason for the request and requesting reprocessing of the corresponding event,we propose a novel complaint text classification model based on character-level convolutional network for the issue of complaint data containing strong dissatisfaction and many grammatical errors.Experiments demonstrate that our model can achieve state-of-the-art results on Chinese and English complaint texts.2.Express delay prediction:Express delay prediction is a very important indicator of express delivery service,and nearly 40%of the many reasons for user complaints are related to delays.Delay analysis plays a very important role in reducing user complaint rates.At the same time,considering the large scale of postal data,this study proposes an express delay prediction model based on express status data to solve the this problem.3.Name rationality analysis:The post office responded to the demand for the real name system of the country and has gradually adopted the real name system into the delivery process.The complaint department will soon adopt the real-name system.Before accessing the public security data,this study has developed a name plausibility check rule for preliminary determination about whether it is a real name.4.Data analysis system for express complaint business:The above three research points are based on algorithms or rules analyzing the complaint text,and the analysis results can be used as the theoretical basis of the complaint business data analysis system.In order to reduce the difficulty of analyzing those user data related to the complaint business and improve the efficiency of complaint processing,this research point comprehensively considers the needs of the complaint users,and conducts research and implementation of the complaint business data analysis system based on the existing parallel computing framework.