Abstract. Because most schools have been using traditional methods to manage students, there is a lack of effective monitoring of students' behavioral problems. In order to solve this problem, this paper analyses the characteristics of big data in University campus, adopts K-Means algorithm, a traditional clustering analysis algorithm, and proposes an early warning system of College Students' behavior based on Internet of Things and big data environment under the mainstream Hadoop open source platform. The system excavates and analyses the potential connections in the massive data of these campuses, studies the characteristics of students' behavior, analyses the law of students' behavior, and clusters the categories of students' behavior. It can provide students, colleges, schools and logistics management departments with multi-dimensional behavior analysis and prediction, early warning and safety control of students' behavior, realize the informatization of students' management means, improve the scientific level of students' education management, and promote the construction of intelligent digital campus.
Cooperative Innovation Behavior Based on Big Data