Research And Application Of Bayesian Network In The Knowledge Map

With the rapid development of the Internet, e-learning has drawn increasingly attention from the public, more and more learners choose to study through the Internet. Due to the explosive growth of online learning resources, online learners are easy to encounter the "learning disorientation" and "cognitive overload" when they are learning on the Internet. At the same time, the present online education continues to focus on the problem of personalized learning and differential learning. Online education has not been well developed to promote educational equity and improve the efficiency of learning. In this thesis, Bayesian network theory and technology is applied to the learning process of online education. This thesis has built a knowledge map model for online education and proposed a mastering degree of knowledge prediction method based on Bayesian networks inference mechanism to provide personalized learning navigation for students.The main work of this thesis includes:1. This thesis analyzed the organizational characteristics of subject knowledge, mapped the knowledge units and their mutual dependencies into nodes and edges of a knowledge network, determined the node attributes of knowledge units, constructed subject-oriented knowledge map model, researched the calculation method of the centrality and difficulty of knowledge nodes, and on the base of these study, presented an efficient method to generate learning path.2. This thesis regarded the subject knowledge as the structure of Bayesian network, combined it with the evaluation results of item response theory to construct Gaussian Bayesian network model that can quantitatively calculate the degree of dependence among the knowledge units, and thus put forward a prediction method of the knowledge unit master degree, made using computers prediction results to help learners control and adjust the learning progress is possible.3. This thesis designed and implemented the online learning system based on Bayesian forecasting mechanism. After learners studying the relevant knowledge units and completing the exercises, the system can give learners the next appropriate knowledge point to study and provide learners with a personalized learning path navigation.Knowledge map model proposed in this thesis can vividly reflect the relationship of disciplinary domain knowledge, gives the optimal learning paths to avoid the appearance of “learning disorientation” and “cognitive overload” effectively. After experimental analysis, Gaussian Bayesian network model proposed in this thesis can effectively reflect the degree of dependence among the knowledge units. By comparing and analyzing the predicted value and the actual value, the result shows that prediction method of the mastery of knowledge proposed in this thesis has relatively high accuracy which can provide individualized learning paths navigation for students and has a good application prospect and value.