Research On MOOC Recommendation Algorithm Based On Content And Word2vec

Since the 18 th National Congress of the Communist Party of China,the state has attached great importance to the development of Vocational education.Secondary vocational education is facing unprecedented opportunities for development.In order to promote the teaching reform of vocational education and improve the quality of personnel training in vocational education,the Ministry of Education proposed to construct the teaching resource library of vocational education.The construction of teaching resources library of vocational education is an important means to promote the comprehensive application of information technology in the teaching reform and teaching implementation of secondary vocational education in accordance with the trend of "Internet +".As an important part of the teaching resource library of vocational education,MOOC platforms can share more high-quality teaching resources of secondary vocational education,which provide favorable conditions for secondary vocational education learners to carry out personalized learning and lifelong learning.However,with the construction and development of MOOC platforms,information overloading will become increasingly highlighted.Faced with a wide variety of online courses,learners often need to spend a lot of time to screen for relevant courses.At present,the major MOOC platforms provide only the classification and search function of the courses,which is not intelligent and humanized enough in course presentation.Facing rich courses resources,how to help the learners to locate the target course quickly to promote individual learning and to provide a supportive environment for learners to learn intelligently,is also an important issue to be considered in the process of building intelligent MOOC platforms and making better use of teaching resources library of vocational education.In this paper,we took the "www.icourses.cn" web-based MOOC platform as an example and customized the Scrapy crawler framework to crawl the course data first,and then applied the natural language processing technology to preprocess the crawled text data.Then,we proposed a content-based and word2vec-based course recommendation algorithm to improve the defects of traditional content-based recommendation algorithm in semantic analysis of object modeling and similarity calculation.At last,we verified the validity of the model by the experiments on the crawled course data set.Two innovative points are presented in this paper:1.Application innovation: Recommendation algorithm is applied to MOOC to extend the application of recommendation system.2.Algorithm innovation: The proposed improved recommendation algorithm combines TF-IDF weighting algorithm with word2 vec word vector to represent items,so as to improve the semantic analysis ability of the traditional content-based recommendation algorithm;A new method of similarity calculation is proposed,which weights the texts similarity based on TF-IDF algorithm and word2 vec word vector,taking account of the words frequency and semantic information of the texts comprehensively.