Analysis and Visualization of Students’ Learning Based on Multi-Topic Chat Text

The popularization of online chatrooms facilitates students to discuss, question and answer based on their learning content. The chat text reflects students' learning concerns, learning progress, learning enthusiasm and so on, which has important reference and analysis value. But the previous studies on students' chat text were mainly to explore the participation behavior and interaction patterns, while less attention was paid to the dynamic process of students' learning. Meanwhile, the existing technology lacked effective analysis methods for the highly variable and multi-topic commingled chat text. Therefore, this paper proposes the Multi-topic Analysis by Extraction and Combination (MtAEC) architecture. On one hand, MtAEC locks the questioner and respondent in the chat, tracks and collects their text, and calculates the text similarity to verify whether the lock result is correct. This method confirms the role of the questioner and the respondent in the chat. On the other hand, to solve the problem of topic analysis in the multi-topic commingled chat text, MtAEC points the parent pointer to the interrogative sentence, and traverses down the child pointer to search the answer. In addition, this paper analyzes the topic of the students' chat and visualizes the experimental results in the form of charts based on the group chat text of a real course. Then, we summarize the students' learning concerns, learning progress and learning enthusiasm, which are helpful for teachers to better understand the dynamic changes of the students' learning process.