Research On Knowledge Resource On-demand And Service Composition Method In E-learning System

E-learning is a teaching style which is web-based, electronic, digital and multimedia. In practical applications, the scope of the E-learning is very broad. It includes not only network learning based on internet, but also digital learning based on multimedia. Knowledge learning of E-learning in this paper is a gradual progress. New knowledge is dependent on the knowledge that learners have mastered.There are some preliminary studies on personalized, adaptive and collaborative E-learning systems. These studies are based on the learner’s personal learning background. Learning content is studied with text mining method and there is no combination of the characteristics of E-learning in specific areas. Although E-learning is convenient and efficient, because of the load of the E-learning environment, there are still many shortcomings and problems. Firstly. because E-learning system provides huge resource, learners often tend to feel disorientation when they use the system to learn. They cannot make their own learning plans effectively. Secondly, E-learning system is involved in the course website. teaching materials and many other data sources, so the effective integration of resources in a major challenge. Thirdly, E-learning system should provide services to meet the individual needs based on the learner’s foundation and interests. In order to solve these problems, researches on the on-demand of knowledge resources and service composition in E-learning system is done in this paper. It also supports the dependencies self-recognition among knowledge resources.At the beginning, based on machine learning and natural language understanding technology, a knowledge map automatically generated algorithm is designed in this paper. Knowledge map is a digraph. Dependencies of all courses are the edges of this digraph. Discovering Knowledge unit algorithm is designed according to the organization of the teaching materials. Then the dependent relationship among courses is determined by using the reference among knowledge units. And the knowledge map can be updated dynamically according to the update, additions and deletions of the course materials. What’s more, the service composition model of multiple data sources in this E-learning system is designed using workflow model. In this way, service modeling, selection and implementation process are standardized. Service discovery and selection are completed by workflow engine. This method is dynamic and easy to monitor and manage the running process. Finally, the traditional collaborative filtering recommendation algorithm doesn’t consider the user interest changes and user’s background. In response to these flaws, an effective recommendation algorithm is proposed. Based on traditional collaborative filtering algorithm, cognitive function and correlation function are introduced; effect of time and user preference is considered. Then the on-demand of knowledge resource in E-learning system is supported and individual needs of the user are met.Based on the above, this paper realizes a SaaS-oriented E-learning system and improves the using process and service life cycle management. The system is able to automatically identify the dependencies between the courses and meet the user’s demand for knowledge discover and customizing. Workflow engine is used in this paper to organize the resources to provide services to users. This also facilitates administrators to monitor the system. When learners use the system, they are able to quickly customize knowledge resources which meet their personal demands of the learning objectives. In this way, the efficiency of learners can be improved and their intended learning goals can be achieved.