Font Size: a A A

Research On The Personalized Advertising Push Services For Internet Users

Posted on:2015-01-11Degree:MasterType:Thesis
Country:ChinaCandidate:M ZhouFull Text:PDF
GTID:2268330425481828Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In recent years, with the rapid development of broadband Internet and the number of Internet users continues to rise, advertising push services on the Internet ushered in new opportunities. Advertising push services have been a large number of portals, shopping sites and social networking sites and other concerns. In Internet advertising push service, the site of its advertising space provided by way of compensation for the use of advertising to advertisers. When advertisers use push services, on the one hand want to improve the effect of product advertising push, timely and accurate forthcoming advertising push to be interested in these products target customers; on the other hand, also hopes to expand the scope of advertising push, is about advertising pushed to more target customers. To this purpose, the emergence of new advertising push by the operating mode of a data management platform (DMP) and demand-side platform (DSP) composed. In this operating mode, data management platform is mainly responsible parties to collect user data, and user data collected to analyze characteristics of users finally get interest; DSP DMP will feature user data as data entry, advertising main customize their bidding rules, and ultimately by the DSP platform operators will be pushed to the user AD, the emergence of new advertising push by the operating mode of a data management platform (DMP) and demand-side platform (DSP) composed. In this operating mode, data management platform is mainly responsible parties to collect user data, and user data collected to analyze characteristics of users finally get interest; DSP DMP will feature user data as data entry, advertising main customize their bidding rules, and ultimately by the DSP platform operators will be pushed to the user AD.Internet service provider (ISP) for the provision of Internet service providers, as opposed to a single electronic portal, which covers the number of users is much larger, and its user base is widely distributed. When a user in the use of Internet access services provided by the ISP, the ISP logs leaving a record of the user access to all sites. The log data contains a more comprehensive user interest features. Seen through the analysis of user interest in ISP’s log feature data has a unique advantage in the Internet advertising push service applications, it has become an important source of data is the DMP.The topics to users of broadband access to records of a carrier-based data management platform in the data analysis of the data obtained on the basis of user interest feature library, on the basis of the use of information push technology, personalized advertising push technology, design and implements personalized ads based MapReduce parallel framework push service system.Firstly, for Internet users to personalize advertising push demand service description and design of the overall architecture of personalized advertising push service system. Personalized advertising push services data input data management platform also conducted in-depth exposition, which aimed to identify valid cookie entries aggregation session, eventually reaching the log file to identify individual users.Then, we consider MapReduce on a large data processing has the advantage of trying to MapReduce with the traditional method of combining real-time bidding, a distributed real-time bidding algorithm, and then applied to personalized advertising push service. In this paper, a Hadoop cluster framework, the design of distributed algorithms for efficient MR-RTB advertisers’ real-time bidding algorithms, and apply it to the advertising push service. MR-RTB core algorithm design efficient data partitioning strategy and join query processing algorithms added frequently, the subject of the use of Hash function avoids data partitioning strategy and exit nodes. It greatly reduce the average response time limit.Finally, the subject of the use of appropriate open source Hadoop platform and Eclipse development environment on advertising push service for encoding tests. Tested in real advertising push service can be a good combination of advertisers, users, operators demand advertising tripartite advertising push. Ultimately meet the individual needs of users; advertisers increase their CTR to expand its user base; effectively improve the broadband operator’s advertising revenue.
Keywords/Search Tags:Advertisement Push Services, Real-Time Bidding, AD Precise Push, Hadoop, MR-RTB Algorithm
PDF Full Text Request
Related items