Author(s)

Xinyi Gong

Corresponding Author

Xinyi Gong

Abstract

School bullying has become a very important problem in education circle. In order to discuss this problem in the rational and scientific eye, we can make use of big data analysis that can provide data and decision support. Aiming at the time-varying abstract characteristics, this paper proposes an improved algorithm on school bullying based on K-means clustering, establishes the whole data analysis structure of data collection, data processing and inferential decision, analyzes the data feature of school bullying. It also gains the time, geographic and age characteristic though chi square test, linear regression and logistic regression.On the other hand, behavior data is also processed with K-means clustering algorithm. Time varying characteristic and uncertain factors are greatly reduced through density, grid and models. In the end, data support for the solution of school bullying behavior is offered by means of inferential decision based on neural network.

Keywords

Component, Formatting, Style, Styling, Insert.