Compared with multimedia teaching, E-learning is a brand-new teaching mode, and its emergence is another wave in the development of the Internet. E-Learning will enable better development space for teaching resources, teaching management, and teacher-student exchanges. The concept of E-learning is constantly being given new meaning, and various teaching methods based on the E-learning model are also changing from auxiliary teaching methods to major teaching models. The purpose of this study is to explore the direction of activation of e-learning activity model in order to induce college students to develop knowledge expertise and critical thinking ability. For this purpose, it is possible to predict the learners learning style, learning progress and learning participation by analyzing the accumulated information from the e-learning activities using educational data mining. This will help to build an effective learning model and to help students understand the various uses of learning analysis to provide customized education tailored to the characteristics of individual learners. In order to calculate the frequency of learning activities of COLLA system learners, the top three groups actively pursuing learning activities based on the Group Activity Index and the three subgroups in which learning activities were not active were classified and analyzed. Finally, the direction of e-learning model development based on educational data mining is suggested.