Research And Implementation Of Precision Advertising Based On Behavioral Targeting

With the progress of science and technology and the improvement of living standards, Internet users increased significantly. According to the IDC report said the world in 2012 the number of Internet users will exceed 1.9 billion, the number of Internet users in China will exceed 500 million.Such user scale, so that all enterprises tend to rely on online advertising to promote their products and services. At present, the current Internet advertising has such disadvantage:no target, Explosive launch, can’t meet the extension of the intended. Precision placement on advertising study it in depth.Currently, Precision of the online advertising delivery, there are three main ways:based on the physical location of the classified ads; based on user behavior, content targeted advertising; based on mobile terminal SMS advertising, which targeted advertising coverage, high accuracy, is the major Internet companies and the precision of the advertising service provider of the pie. The main subject of research is based on the precision of user behavior targeted advertising, and for the design and implementation. This paper first introduces some background knowledge and precision of the delivery status of technology, describes the system after the implementation of the relevant terminology, technology, and the use of classification algorithms, focusing on system design and implementation process, including the database design and design process of each module.In the optimization section, proposed an optimization algorithm, is about the user’s behavior is divided into short-term and long-term behavior, online advertising tracking through the window on the user’s web browser to track. Based on user online behavior of Regency, Intensity, Frequently value and weight of the system requirements, respectively, long-term behavior and short-term user behavior profiling, dig out the preferences of the user’s short-term and long-term preferences and calculate the Raw Score. According to long-term behavior and short-term behavior of the weight and Raw Score, to calculate the user’s score matrix of each subject, and accordingly advertising, experimental data show that optimized module processing, advertising a certain accuracy improvement, this module the innovation of the entire system.At last, concludes the paper; proposed shortcomings of the system; made the focus of future work. Summarizing the work and outcomes during graduate.