A Combined Approach for Search of Learning Objects on the Web

The present thesis describes a system for search and delivery of learning objects. This system combines different methods of search: keyword- and concept-based search and personalization. The keyword- and concept-based search methods are responsible for establishing the relevance of each learning object to the query. The personalized search of learning objects calculates relevance of each document based on a comparison of the learner (user) profile and learning object descriptions. Such a comparison is based not only on the values of characteristics of the learner profile and attributes of the learning object descriptions, but also on the importance of these characteristics and attributes for the learner. These approaches are combined in our integrated System of Delivery of Learning Objects (SDLO). The relevance of learning objects to the learner's query is determined using a combination of these relevance measures. The effectiveness of the proposed approach is evaluated, including the determination of the relevant optimal weighting coefficients.