| In recent years,online education platforms represented by MOOC platforms are developing rapidly.Online education has no thresholds such as high school and college entrance examinations,and there are no requirements and drawbacks that must be studied in a certain place.It makes everyone able to enjoy the best universities’ education equally,and learners have rights to choose the courses that are fashionable,or they want to learn and are interested in freely.However,while this provides extremely high convenience for learners in learning,it also separates the direct contact between learners and other people in the traditional classroom learning environment.As a result,learners are always accustomed to studying alone,and have no motivation to actively contact or are not accustomed to contacting indirect learning partners on the Internet.The lack of learning partners to communicate and study together results in low learning efficiency,low learning persistence,high course dropout rates,and so on.This paper examines the key elements of traditional pedagogy for the construction of learning partners,combines the user behavior characteristics of online education platforms,and refers to the process of traditional e-commerce recommendation systems to construct a set of online education learners’ collaborative learning social relationship construction methods.And preliminary exploration of the engineering design of the learner’s online education interactive platform.This article conducts research on four sub-problems on how to solve the topic recommended by learner partnerships on online education platforms,namely,the feature extraction of learners on the online education platform based on educational psychology,the construction of learner’s basic partner network and the discovery of learning communities,and Neural network sorting online education platform learners learn partner matching and partner network simulation convergence based on scale-free network theory mixed with partner recommendation lists.For the learner feature extraction sub-problem,this article first explores the similarity,proximity,compensation,appearance and other important characteristics in the process of building a good learning partnership between learners in traditional pedagogy.Next,it analyzes the differences and unique characteristics of learners’ behavior characteristics between online education platforms and traditional education platforms.According to these characteristics,it is targeted and selective to strengthen some of the influencing characteristics in the traditional education environment,and also proposes some unique characteristics,weakens or even deletes some unobvious characteristics of other online education platforms.It is given.Online education platform learning partners recommend specific quantitative methods for key factors.For learners’ basic partner network construction and community discovery subproblems,this article is based on the user interaction graph theory in traditional social network research,combined with the scarcity of direct interaction among learners on online education platforms,and based on the above-mentioned research on the key elements of online education platforms.The platform-specific learning algorithm for basic interactive network construction of learners.After that,according to the user behavior characteristics of the online education platform,three different network partitioning community algorithms were targeted to improve and experiment,namely the kMean+Jaccard algorithm,the Girvan-Newman interface cutting algorithm,and the random walk algorithm.And finally choose the edge cutting algorithm for engineering practice and further experiments.For the neural network ranking sub-problem,this paper firstly proposes an online education platform partnership random annotation algorithm based on the traditional partner annotation method in the field of education,combined with the characteristics of the online education platform.This method is combined with time series segmentation to generate neural network training set and test set.Next,according to the different learner features extracted before,the different features are vectorized with a targeted network structure and combined to generate the final neural network input.And designed the final overall structure of partner classification network,and completed the experiment of learner partnership matching in the previously divided learner community.At the same time,a qualitative analysis of the influence of different learners’ characteristics on the construction of partnerships was carried out.For the sub-problems of network convergence and partner recommendation list mixing,this paper first analyzes the scale-free network structure characteristics of the network structure formed between learners and their recommendation partners on the online education platform over time,and then analyzes the scale-free network structure.The theory of network generation gives the network simulation evolutionary convergence algorithm for learners and their partners in this subject.Next,starting from the social triangle theory,a recommended partner mix-and-arrangement algorithm is proposed to solve the "cocoon house effect" of learner communication after the partner network converges.According to experiments,it is proved that the current mixed-row recommended partner list has better accuracy than the original recommended partner list. |
