| With the rapid development of science and Internet,which provides environment support for e-learning,online learning has been widely spread.At the same time,COVID-19 break out in a wide range of the world in the begin 2020.Online learning can ensure educational institutions can carry on teaching activities normally with the advantages of breaking the time and space restrictions.On the one hand,online learning has great advantages in the sharing of educational resources that make some poor and backward areas also enjoy the courses of famous teachers.Online learning has played a great role in the better distribution of educational resources in China.More and more learners begin to learn from the Internet,which produces a large number of online learning behavior data.In the process of online learning,data in some way could reflect the state of learners in the learning process.These data are more convenient to collect and deal than traditional learning data.And to some extent,it can reduce the error of human factors.Based on the formerly research,this paper explores the relationship between the learning behavior of college students and their learning effect in the autonomous learning environment of network courses which from the main courses of college students.Specifically,the research work is based on the traditional principal component analysis and hierarchical clustering analysis,combined with the support vector machine and Ada Boost strong classification method in machine learning to study the network learning behavior of college students..The results are divided into two aspects:(1)From the data of principal component analysis,there is a significant correlation between the final course performance and classroom performance.Generally speaking,the college students who have a good performance in their usual online learning behavior will get a higher score in their final exam;On the contrary,the student who have a poor performance in their usual online learning behavior will get a lower score in their final exam.(2)Based on the results of traditional cluster analysis and AdaBoost,we can divide college students into two groups.The first group has a high degree of activity and participation in the six types of online learning activities which is higher than another group.Compared with other types of college students,this kind of college students can make full use of the rich learning resources and the convenience of access of the Internet to learn the incomprehensible places in class and preview the difficult compulsory courses in advance.This kind of College Students’ learners are the heart in online learning activities.They pay more attention to their own performance in the process of online course learning and autonomous learning activities,and have more prominent behavioral characteristics.In some way,their high enthusiasm and participation in online learning activities can inspire and guide other college students’ participation and engagement.Therefore,we can define this kind of college students as "high-level participation and engagement learners".The second kind of College Students’ learners have a lower participation and engagement in e-learning activities,which means that this part students do not pay much attention to learning performance and course performance.They may participate in the course learning of supernova in order to complete the platform learning tasks assigned by teachers or not be named by teacher.Some college students performance may also not very well in the whole learning behavior due to the lack of supervision from teachers and their poor self-control ability.Their learning behavior is not outstanding.And they show the phenomenon of poor overall enthusiasm and low participation in learning.Therefore,we can define this kind of college students as "low-level participation and engagement learners".In addition,this kind of College Students’ learners account for about 70% of the total learners,which indicates that most of the college students’ learners are in a passive and negative learning state in the network learning activities.When they return to school to study the next course,their academic performance may fall behind,and even lead to a certain probability of course failure. |
