Date of Award
Master of Science (MS)
Content-targeted advertising system is becoming an increasingly important part of the funding source of free web services. Highly efficient content analysis is the pivotal key of such a system. This project aims to establish a content analysis engine involving fuzzy logic that is able to automatically analyze real user-posted Web documents such as blog entries. Based on the analysis result, the system matches and retrieves the most appropriate Web advertisements. The focus and complexity is on how to better estimate and acquire the keywords that represent a given Web document. Fuzzy Web mining concept will be applied to synthetically consider multiple factors of Web content. A Fuzzy Ranking System is established based on certain fuzzy (and some crisp) rules, fuzzy sets, and membership functions to get the best candidate keywords. Once it is has obtained the keywords, the system will retrieve corresponding advertisements from certain providers through Web services as matched advertisements, similarly to retrieving a products list from Amazon.com. In 87% of the cases, the results of this system can match the accuracy of the Google Adwords system. Furthermore, this expandable system will also be a solid base for further research and development on this topic.
Wang, Yezhou, "Fuzzy Content Mining for Targeted Advertisement" (2008). Master's Projects. Paper 117.