MOOC Learner's Interest Modeling And Case Analysis

As a new method of online learning,because of its openness,fragmentation of curriculum organization,and interactive features,MOOC has attracted attention from educational institutions and learners at home and abroad since its inception.Researchers have carried out various researches on data mining,natural language processing,and information retrieval based on MOOC,such as educational data mining and learning behavior analysis.Currently,one of the most important directions for MOOC study is to meet the personalized service needs of learners.One of the key issues in personalized services is how to model the interests of learners.The learner's interest modeling in MOOC can reflect the learner's preference in an accurate and real-time interest,which is the basis for personalized service,and directly affects the quality of the MOOC personalized service.This article mainly focuses on the MOOC learners interest modeling technology.The main research results include:(1)Comprehensive construction of the interest model of learners in the MOOC environment,namely(S,C,T)triples.Where,S stands for the collection of subject interest classes,C represents the collection of courses interest classes,and T represents the collection of topic interest classes.The model can reflect learners'interest topic and their weights to a certain extent.(2)The learners' interest in quantitative breadth and intensity of interest in two indicators,from the three dimensions of subject,courses and topic,the interest of learners is calculated.Among them,according to the division of the subject categories on the MOOC website,the learners' interest breadth of the subject level is obtained from the statistics of the subject of the learner's course.The interest intensity value of the discipline level can be obtained by calculating the proportion of each subject category.The breadth of interest of learners at the level of the course is derived from the courses which are involved in the course certificate and course review.The calculation of interest intensity uses the multiple linear regression method,and the linear regression equation established by the two indexes of the learning certificate and the number of comments can obtain the interest intensity of the course level.The breadth of interest in the topic level is extracted by the tf-idf algorithm,to extract the keywords of each course,and the Angle cosine value is calculated for the semantic similarity of the keywords,and different keywords are obtained by clustering.The calculation of interest intensity is the key word and its weight.Finally,some cases is given to analyze the interest model.This paper selected the Chinese University MOOC website learners as the research object,and grabbed 21 active learner homepage explicit information data,including the course number,certificate,the course review information,based on the data from three dimensions to build the subject.course.topic interest model.Through the analysis of the learners' interest in model visualization,show that the model can more comprehensive introduction to the learner's"subject-course-topic" interest set,which lays a good foundation for the following personalized service..