Education data mining: How to mine interactive text in MOOCs using natural language process

This paper proposes a combination of data mining and natural language processing technology, try to analyze students' learning behavior and content in MOOCs interactive part, to dig their learning interest, difficulty, tendencies, to evaluate their homework effect, through the interaction between teachers and students, students posting, homework or answer content, preventing of cheating behavior, and so on. So that, teachers optimize the course structure, establish of incentive mechanisms to stimulate students' learning desire, improve the completion rate of the courses. This method can dig valuable information from the seemingly random mass data, optimize the corresponding key elements in the teaching model, so that, teaching works go in a virtuous circle.