Research of Analyzing MOOCs’ Learning Records with Neural Network

• Main Authors: Chi-LinChen, 陳麒麟
• Other Authors: Chu-Sing Yang
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/77awpt

碩士 === 國立成功大學 === 電腦與通信工程研究所 === 106 === With the development of technology, the learning methods are different from the past. MOOCs(Massive Open Online Courses) platform like Khan Academy, Cousera and edX are more and more popular from 2012. These platforms contain amount of user’s data and researchers want to know what kind of knowledge we could know from these data. However, it is difficult to analyze data from different learning system together. Moreover, current learning analytic service on platform is not smart enough. Therefore, we develop a system architecture to collect learning records from different systems and implement a LSTM neural network to predict students’ performance. And the accuracy of prediction is higher than 75% normally and the error rate is lower than 10%. Also, the system could transform the relation between learning material to relation graph. It will be helpful for teacher to adjust their teaching strategy.