BClassifier: a personal agent for bookmark classification

For the study of software systems helping users reduce personal tasks, we implemented an agent, called BClassifier, which classifies bookmarks on behalf of the user. The BClassifier is an agent based on a naive Bayesian learning algorithm which acquires knowledge through training with examples. We present the findings of experimentation where the BClassifier consistently exhibits success rates over 70%. Two other types of learning methods are compared with the naive Bayesian technique