This paper analyzes the influences on creative thinking by utilizing multivariate data integration methods and modular network-based integration methods. A creative drive gene identification algorithm was built using the network learning approach. Using the EMD method, a subset of creative thinking genes is obtained, and the histogram data density of each gene expression value is calculated. Creative thinking was classified using the SVM classification algorithm. Games have been shown to promote creative thinking in elementary school students to some extent, according to the results. The personalities of 60% of the students change, and 50% come up with new ideas during the games.