Topic Mining Methods For Massive Open Online Courses

With the development of communication technology, the emergence of the popularity of mobile devices and cloud computing technologies in recent years, the data in the Internet shows explosive growth, information technology is entering big data era. Data mining technology is one of the hot spot in big data area, how to effectively use data mining technology in massive open online courses is the main purpose of this thesis.Massive open online courses, also known as MOOC, is a new educational model emerging in recent years, which is a free online class open to the public. As more and more people paying attention to MOOC, researches in online education are also start up. Research achievements in MOOC can improve the quality of teaching and bring best educational resources to all corners of the world.The main idea of this thesis is to use data mining techniques to analyze MOOC data, from discovering students' learning pattern through online learning data to provide a scientific basis for improving the quality of online education. In this study, data collected from Coursera platform are used for data mining as follows: 1) C lean and preprocess the data, then extract users' profile feature and use LDA model to get topic feature; 2) Combine the users' profile feature and topic feature to classify threads in MOOC discussion forums; 3) Use users' text feature to improve collaborative filtering algorithm and recommend threads to users in MOOC discussion forums. After analysis and verification, it is proved that the model used in this thesis can accurately classify the threads in MOOC discussion forums, in the mean time the algorithm can recommend threads to users depends on their interests properly. The proposed method s in this thesis provide a scientific basis for improving the quality of online education, with some guide and refer effect.