Medical Advertising Content Filtering System in Online Knowledge Sharing Service

Abstract

Online knowledge sharing service such as Q&A communities in the Web is a representative service as collective intelligence among people and a critical web-service to entice online consumers as well. A huge volume of Q&A content is generated and shared in real time, however there are also uncertain information including not only valuable knowledge but also commercial data such as advertisements, marketing, and even wrong information. That commercial content is able to lead online-users to controversy. This study proposes a content filtering system in the knowledge sharing community to classify whether the information is commercial or not. The filtering system applies linguistic feature sets and employs a support vector machines algorithm in machine learning methods to classify whether the information in the knowledge sharing community is commercial or not. To build the algorithm and validate the system, we set the target domain within the healthcare content and gathered question and answer content about lung cancer of knowledge sharing service in a Korean web-portal site, Naver.com. As the result, the proposed system accomplished accuracy average 84.0% with the Ads words and the document length.