A Proposal of Discovering User Interest by Support Vector Machine and Decision Tree on Document Classification

User interest is one of personal traits attracting researchers\´ attention in user modeling and user profiling. User interest competes with user knowledge to become the most important characteristic in user model. Adaptive systems need to know user interests so that provide adaptation to user. For example, adaptive learning systems tailor learning materials (lesson, example, exercise, test...) to user interests. We propose a new approach for discovering user interest based on document classification. The basic idea is to consider user interests as classes of documents. The process of classifying documents is also the process of discovering user interests. There are two new points of view: 1. The series of user access in his/her history are modeled as documents. So user is referred indirectly to as "document". 2. User interests are classes such documents are belong to.