Innovative and entrepreneurial characteristics of university students based on logistic regression model

Abstract

This paper proposes a logistic regression model with structural sparsity to study the characteristics of innovation and entrepreneurship among college students. The article first analyzes the basic form of the logistic regression model, including the objective function and the selection method for the penalty function. Then, because the ADMM algorithm combines the advantages of augmented Lagrangian and pairwise decomposition, which can reduce the computational difficulty and complexity, based on this advantage, this paper designs the ADMM algorithm solution framework that is favorable for distributed computing. Finally, this paper analyzes the relationship between the development of innovation and entrepreneurship ability of students in R colleges and universities and their gender, grade, academic foundation, experience in clubs and discipline type. The results yielded that college students’ mean value of innovation and entrepreneurship competence in HEI R was 3.734. The mean value of the scores of each sub-competence ranged from 3.531 to 3.918, which puts the overall innovation and entrepreneurship competence of students in this university at an intermediate level. Therefore, this study plays an important role in understanding the innovation and entrepreneurial characteristics of students in higher education.

Keywords

• logistic regression
• admm algorithm
• augmented lagrangian
• dyadic decomposition
• innovation and entrepreneurship characteristics
• 97q70