ABSTRACT
Using special technology, website-visitors’ behaviors could be thoroughly observed, and their behavioral characteristics can also be accurately checked. Additionally, visitors’ demand could be directly reflected on the basis of visitors’ behavioral characteristics. Finally, the targeted advertisements are sent to these visitors according to their needs and preference. However, it is really difficult for accurate targeting on the customers. The specific policy of behavior-targeted advertising is formulated by the analysis of various factors including time, space, the visitors’ information and their needs etc., which is aimed to provide correlated advertisement information using on-line browse activity. Then, visitors’ valuable behavior information is excavated by deep analysis of webpage historical access [2]. In this situation, users’ browse behavior needs to be classified, and clustering algorithm can be used to seek meaningful clusters, which induced similar results in homogeneous
1 INTRODUCTION
The online advertising is rapidly developed and increasingly becoming the essential part of advertising industry due to the popularity of Internet. Moreover, the online advertising is gradually becoming the critical part of social media marketing strategy for the enterprise because of its special advantages including real-time interactivity, extensive spread boundary as well as personalized service etc. when compared to traditional four communications media including newspaper, magazine, TV and broadcast. According to a previous report of Interactive Advertising Bureau (IAB), the past decade has witnessed the rapid growth of online advertising with an annual rate of 21.7%. In 2008, the total scale of network advertising has rapidly increased to 25.8 billion dollars. Meanwhile, the online advertising in network market also exceeds over 17 billion RMB. Therefore, according to the data from JPMorgan Chase & Co., NYSE, Data Center of China Internet (DCCI) predicted that a high-speed growth tendency of global online advertising will still be kept on in the next few years at a rate of 30%. In China, the scales of online advertising has already achieved to 2.88 billion RMB in 2010, still the figure will promptly rise due to the enormous amount of net citizens especially those who use cell phone to browse webpage at a
observation and different results in inhomogeneous observations. The classification of users is involved in various factors, and the classification indices and weight of each factor are often different, thereby reflecting the vague of classification [4]. Therefore, in this article, we proposed a mathematical model of fuzzy clustering analysis with multi-indices and multilayer on the characteristics of users’ browse behavior. Only in this way, the targeted customers could be well classified and the targeted-advertising could be more accurate and effective.
