Research on Application of Frequent Pattern Growth Algorithm in Academic Early Warning

Abstract

The arrival of big data era has led to imperative management changes in the field of education. A large number of chaotic and ineffectively used of education big data exists in the process of intelligent campus construction. By combining big data technology with educational administration to improve students' learning skill and strengthen the construction of high-quality style of study. For example, this paper takes the final grades of 582 students of mathematics in a university from grade 2006 to grade 2015. Firstly, to eliminate the different impact of different teachers on the test paper scoring standards, we standardize the scores of all students. Secondly, we code the scores of each student using the improved Frequent Pattern Growth algorithm to the correlation between the failed courses (for convenience, we will abbreviate the Frequent Pattern as FP.). Then we find out the correlation degree between the failing courses by the jackknife method. Finally, for improving the academic early warning system, we verify the feasibility of the improved FP Growth algorithm for early warning of failing courses using the data of mathematics students majoring in grade 2016.