Design And Implementation Of The Quality Control System For Native Advertising

In the context of the wave of big data swept across the globe,traditional Internet advertising has become increasingly inefficient as audience media literacy continues to improve and the mobile Internet has developed rapidly.In order to achieve new development,the traditional Internet advertising industry urgently needs a new form of advertising to break through the dilemma.As a result,native advertising came into being.Native ads are a profit model based on the experience of websites and App users driven by advertising content and user interest.Native ads present brand information through "harmonious" content,without destroying the user's experience,providing valuable information to users.Therefore,in order not to undermine the user experience,we need to pay more attention to the quality of native ads.At the same time,advertising as a major business of the company,the company hopes to have a macro understanding of the quality of advertising and control.In this context,the native advertising quality monitoring system came into being.The system is divided into two main parts:advertisement quality inspection,advertisement quality report statistics and display(Part of the system in the system,referred to as ADQ platforms).The statistics of the advertisement quality report and the basic data of the display part depend on the detection results of advertisement quality detection.At the same time,in order to introduce the role of the system in the entire advertising business system,to clarify the source of the data needed in the system,the whereabouts of the output data,and the way and mechanism of data interaction,I separate the relevant introduction of this part as a system.The third part.This article will be introduced from three parts:advertising quality testing,advertising quality report statistics and display(platform)and interaction with external data:(1)The advertising quality inspection adopts Python language and combines the detection techniques of BeautifulSoup,Selenium and PhantomJS technologies to cover the content of advertisement landing pages.Now it has 32 review strategies including vulgar,gambling and category 2 e-commerce.(2)The report statistics and the display part adopt the B/S(Browser/Server,browser/server)structure,based on the company's own ODP development framework,and the front end adopts AngularJS and HTML5 technologies.The platform is divided into four modules:advertising quality,account quality,user feedback monitoring and user rights management.The author participated in the overall requirements analysis,architecture design,database design and interface design of the platform.In the detailed design and implementation phase,the design and development of the four modules were completed,and part of the system's tests were completed during the testing phase.This thesis uses the unified modeling language to describe and analyze the requirements.Through combing the business logic,the architecture design scheme of the platform is proposed,and the process of the author's responsible part is elaborated in detail.(3)The interaction with the external data will introduce the communication mechanism and data interaction between the system and the external system according to two important business flow lines.The system has now completed all the development,in the operational phase,and constantly updated and optimized based on operational results and feedback,and gradually improve the function of the system.The author responsible for the stable operation,basically achieved the desired goal.