Hierarchical Classification for Multiple, Distributed Web Databases

The proliferation of online information resources increases the importance of effective and efficient distributed searching. Our research aims to provide an alternative hierarchical categorization and search capability based on a Bayesian network learning algorithm. Our proposed approach, which is grounded on automatic textual analysis of subject content of online web databases, attempts to address the database selection problem by first classifying web databases into a hierarchy of topic categories. The experimental results reported demonstrate that such a classification approach not only effectively reduces the class search space, but also helps to significantly improve the accuracy of classification performance.