The Research Of Internet Targeted Advertising Systems

The rapid development of Internet, using internet-based targeted advertis- ing systems has been an effective way to post ads. As for current internet-based targeted advertising algorithms, there are some problems,including content matching accuracy is low and customer preferences are not considered properly. To solve these problems, this paper proposes a semantic-cluster based content targeted advertising algorithm. And base on it, considering the history of customer interactions and learning user preferences, this paper proposes a user preferences based behaviours targeted advertising algorithm. By litera- ture review and experiment verification, this paper design an internet targeted advertising system.The work of this passage mainly focuses on the following areas:1. To gain a better content matching between ads and web pages, this paper proposes a semantic-cluster based content targeted advertising algorithm. In the algorithm, the concept of semantic-cluster is defined.Based on this, improved TF-IDF is defined. After extracting key words and relevancy- based semantic expension, the similarity between web pages and advertis- ing content is calculated. Designing experiment and verify the performance of proposed algorithms.2.To make the posting result more accurate, we add user preferences.Based on user daily interaction inforamtion, we learn about individual prefer- ences. With awaiting posting ads, the algoritm calculate user preferences based weight and post the highest to users. Designing experiment and verify the performance of proposed algorithms.3. Give the overall design of the internet targeted advertising system,includ- ing user interaction section, data collection and processing section,tar- geted ads posting section and data storage section. On the basis of improv- ing and perfecting designs, the paper gave the realization of the design.By experiment verification, internet targeted advertising system proposed in this paper achieved expected goals on the improvement of posting accuracy.