Research On Knowledge Base Network Updating Mechanism Based On Markov Decision Process

With the popularization and development of computer,the learning method has gradually shifted from offline to online,and from the same tactics in the past to the learning of thinking mode.Moreover,the implementation of education informatization for many years has accumulated a large number of data foundations.These data have hidden a lot of information that can improve the quality of teaching,and assist in the various difficulties encountered in the decisionmaking education process.With today's huge knowledge system,the topic of ”how to improve learning efficiency” has gradually entered the research scope of scholars in various related disciplines.Many scholars started to study from the personalized learning path,but the results were minimal.In fact,the learning process is not a completely linear process,and everyone 's learning path is different.But the experience encountered in the learning process is for later people,and it is advisable to accumulate these experiences by using appropriate learning content organization methods,which can continuously improve the learning efficiency of related knowledge.Therefore,based on the Markov decision process,this study mines the relationship between learning content from learner behavior data and forms a knowledge base network,so as to organize learning content with a reasonable network structure.The main research content of this article has the following two parts:(1)Learner behavioral data modeling.Analyzing,refining,and organizing the data of the learner's own information,the learning behavior data,and the relevant data of the learning content.Using the network structure to describe the correlation between the data,to form a knowledge base network;(2)Research on network update mechanism.A directed edge in the knowledge base network records the relationship between nodes,the direction of edge constitutes the topology of network,and the relate value of edge describes the relationship between adjacent nodes.As the learning behavior data continuously adjusts the network structure,the knowledge base network evolves.The evolution process in the knowledge base network involves relative position of the nodes and related changes of the edges,that is,the network update mechanism.Moreover,network evolution also has a certain direction.Based on the evaluation mechanism of network superiority,we do some deeply studies on the evolution mechanism of knowledge base network.In order to verify the effectiveness of the network update mechanism,a series of experiments were conducted in this paper to study the parameters of the relate value update formula and the effect of different learning behavior data entry methods on the network superiority.The experiment proves that with the continuous adjustment of the learning behavior data to the network structure,the network superiority continues to decrease and eventually converges.The experimental results further prove that the information carried in the learning behavior data continuously enters the knowledge base network,and the network superiority of the knowledge network converges,indicating that the learning behavior data can effectively guide evolution of the knowledge base network,which is consistent with the theoretical research results.But in this subject,our research is still at a shallower level,and more systematic and scientific results still require a lot of research work.