An Internet Advertisement Service System for Context Classification Based on Support Vector Machine Algorithm

• Main Authors: Chih-Chung Chen, 陳志充
• Other Authors: Hung-Tu Liao
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/88741417419726460685

碩士 === 世新大學 === 資訊管理學研究所(含碩專班) === 97 === 　　The pace of Internet develop starts getting faster and faster in recent years. Internet become a rising medium, it also create an all new market and business for Internet advertisement. For the purpose of increasing Internet advertisement click rate, provide personalization advertisement for Internet user is the most efficiency way. In this thesis, based on Internet user browsing the contents of web page, then the system will provide relative advertisement. Designing a personalization advertising system in this conecept is the goal of this research. 　　The core technology of this thesis use document classification to implment advertising service. This thesis choice the algorithm of “Support Vector Machine” in automatic document categorization to analyze web page. Besides this thesis create new categorization architecture in automatic document categorization. Define “Others categorization” in entire architecture for increasing categorization efficiency. In the meantime, it could solve the complex and rich features in web page. Evaluate effection of categorization. This thesis using Accuracy, Precision, Recall and F-measure Four kinds of measurement methods to estimate. According to testing results, this study could provide extreme match advertisement depend on browsing page and it could assist advertising system to delivery accurately internet advertisement.