Research On Learner's Learning Habits Mining And Learning Path Planning Based On Behavioral Data Analysis

Online education allows everyone to enjoy equal education,and learners can freely choose the courses they want.But it also brings disadvantages,such as the problem of knowledge overload and learning trek caused by fragmentation of knowledge,and the problem of semantic gap between learners and resources in resource recommendation.This paper proposes a method of mining learning habits based on the learner's learning behavior sequence.Using learning habits can remind the learner when to learn,and it can provide the learner with learning path planning and learning content recommendation through analysis of the learning sequence.This is for reducing the dropout rate in online education,or make learners learn as much as possible in the course before the learners gives up the course.At the same time,we consider the use of offline educational resources.In order to let the learner's learning effect be further improved,we incorporate offline learning content with stronger constraints into the planning of the learning path.This paper conducts research on how to solve the two sub-problems in the subject of learning trek and filling the semantic gap between educational resources and learners.The sub-problems are mining the learning habits of learners and planning learning paths for learners.The learning habit mining sub-problem is mainly how to model the learning habits of the learners,and mine the real learning habits of the learners through the processing and analysis of real data.We use the obtained learning habits to divide the learners into groups.Through experiments on real data sets,the effectiveness of using Kmeans thought to solve the problem of grouping learning habits is verified.And through the LSTM algorithm to predict the learner's answer to the problem,it verifies the feasibility of abstracting the learning behavior sequence into a time series for the exercise to predict the answer.The learning path planning sub-problem is mainly how to use the results of learning habits mining to model the learner's learning content and plan a personalized and accurate learning path for the learner,so that the learner can learn more efficiently.For the sub-question of learner learning habits mining,this paper abstracts the learner's learning habits into subspaces in linear space,and obtains the learner's learning habits by processing the real online education dataset Xuetang XDataset.Then,the learners' learning habits are used to divide the learners into learning groups.This paper proposes a grouping method which is based on the Kmeans algorithm and can effectively divide learners with the same learning habits into groups.In addition,learning habits will also be used to predict learner's answer results.This paper abstracts learners' learning behavior into time series,and uses the LSTM algorithm to model each learner and each exercise separately.For the sub-problem of learners' learning path planning,this paper refers to knowledge points and courses as learning objects,and proposes that the learning path is an orderly sequence of knowledge points.We model the knowledge points and courses based on the prediction results of the exercises.Based on the established model,a learning path planning algorithm with review strategy based on knowledge graph is proposed.The algorithm is divided into two parts: in course and between course.The planned learning objects are knowledge points and courses.And we add content related to the integration of online and offline educational resources.Finally,we use topological sorting to show learners a complete personalized learning path.