R & D Intelligent Learning Evaluation and Adaptive Learning Content Recommendation for SOA and Agent-Based e-Learning System by CHOI TOU MAN Thesis Supervisor: Prof. Gong ZhiGuo Thesis Co-supervisor: Prof. Dong MingChui Department of Computer and Information Science As the technologies of computer, network and multimedia have been developed rapidly and applied in wide-area fields, Use these technologies to construct e-Learning system and establish learning community in the society have become reality. E-learning is referred to proceed learning and teaching activities on the network, it is a new mean of learning which is depend on the newly network study environment consisting of Multi-media network learning resources, online learning communities and network technical learning platform. Multi-media courseware create diversifying and interactive learning context comprising not only text, picture, image but also movie, video and audio media, etc. An e-Learner can search for interest information and required courses from network database and conduct study on the basis of individual learning characteristics, interesting, requirement and skill, etc. This educational approach allows learners to conveniently study anytime, anywhere. However, in E-learning system, students and instructors are geographically dispersed, students may feel isolated from the instructors. Instructions are not always available to help the student. Therefore, it is necessary that e-learning system provide learning support services of several aspects include regulation, hint, navigation, motivation, promotion, etc. for students. In the other hand, traditional e-Learning platforms were designed in monolithic way. It has many difficulties and problems in achieving to Scalability, Flexibility, iv Interoperability and Reusability. Due to the reason that the technologies adopted in SOA can be applied in solving these problems, a multi-layered e-Learning system model is proposed and the functions of respective layered is described. Built on SOA, an e-Learning system set up feasible foundation for developing learning support services for e-Learners. For the purpose of providing learning support services, it is need to construct diverse agents to simulate different roles in the teaching and learning environment. Through analyzing the internal structure of agent, multi-agent communication protocol and agent integrated development environment, the method of integrating agents with web services in e-Learning system is designed. Agents integrating with web services is to supply the working mechanism for e-Learning system to develop learning support services. Much web learning courseware is discarded after being used for a temporary period, because the platforms developing the Web learning content and executing the contents are different from each other. Therefore, they could not be used on other platforms. SCORM was developed by the Advanced Development Laboratory (ADL), it is an evolving standard that provides guidelines for developing reusable, searchable, durable and accessible e-learning objects that can be delivered over the Internet to learning management systems. This paper introduces the components of the SCORM, analyzes the hierarchical architecture of curriculum knowledge, and further illustrates the method of designing object-oriented e-Learning courseware based on SCORM. The characteristics of e-Learning determine e-Learning system platform need to provide learning support services for e-Learners. Therefore, it is need to pay attention on researching the method and algorithm for intelligently make learning performance evaluation for e-Learners. This paper demonstrates the purpose of learning evaluation of e-Learning system and introduces the meaning of applying E-Portfolio in proceeding learning evaluation, illustrates the method of setting up summative learning evaluation indexes system, proposes an algorithm to calculate the grade of summative learning evaluation based on Analytic Hierarchy Process. v Because different e-Learner has different prior knowledge, cognitive ability, preferences and learning objective, considering to facilitate adaptive learning , which is referred to dynamically combine learning resources to meet individual e-Learners’ learning characteristics, an intelligent tutoring system embedded in e-Learning system is proposed to derive and recommend learning resource object which the difficulty is appropriate for e-Learners’ cognitive ability for individual e-Learners to enhance their learning effects and increase learning efficiency. The ITS consist of curriculum knowledge organization module, student learning characteristic module and tutoring engine module. In the tutoring engine module, to get rid of the disadvantages of adopting rule set to inference and recommend appropriate learning content object for specific student, an algorithm combining case-based reasoning and item respond theory is proposed to select and deliver appropriate learning courseware according to student’s cognitive ability for individual student. The functions and implement method of each module are described in the paper.