Searching & ranking Learning Objects in a Service Oriented Architecture for e-learning

Abstract

This paper proposes a revised architecture for Service Oriented Architecture (SOA) e-learning system to enhance the reusability of the Learning Objects (LOs) by providing a personalized recommendation model (searching and ranking processes) which is having a shorter processing time, comparing with other approaches. The model counts the similarity degree between the learning objects and the search conditions given by users. The LOs with higher similarity degrees will then be ranked by referring to user's preference history records. Those LOs which are more closed with the user's preference record will be ranked higher in the search result list. The preference record is stored in database to record the details of LOs which are selected by user in the past. As the literature review, some different existing approaches based on recommendation models have been analyzed and compared with each other. Among those existing model, some approaches are focusing on personalized recommendation, while some approaches are focusing on improving the efficiency of the searching and ranking processes. Our contribution is to propose an idea of a recommendation model that is personalized and at the same time, the model should have a shorter processing time. Prototype system will be implemented in future to show the contribution.