Using Machine Learning to Support Resource Quality Assessment: An Adaptive Attribute-Based Approach for Health Information Portals

摘要

Labor-intensity of resource quality assessment is a bottleneck for content management in metadata-driven health information portals. This research proposes an adaptive attribute-based approach to assist informed judgments when assessing the quality of online information resources. It employs intelligent learning techniques to predict values of resource quality attributes based on previous value judgments encoded in resource metadata descriptions. The proposed approach is implemented as an intelligent quality attribute learning component of a portal's content management system. This paper introduces the required machine learning procedures for the implementation of the component. Its prediction performance was evaluated via a series of machine learning experiments, which demonstrated the feasibility and the potential usefulness of the proposed approach.