摘 要
本文研究大学生网络成瘾的影响因素，在调查收集数据前，先确定了影响大学生网络成瘾的因素，即确定自变量，在查阅了相关文献后选择了16种因素作为自变量，在因变量的设定上，考虑到数据的准确行，把原来的网络成瘾是否细分为网瘾严重，一般，轻微，没有；并结合聚类分析中的谱系聚类法来说明因变量取值的合理性，在分析时分别用线性回归分析和用logistic回归分析方法确定造成网瘾的主要因素，并做比较，反应出logistic回归分析的合理性。通过数据分析确定了学习成绩、家庭关系、人际关系、每天上网时间是影响大学生网络成瘾的主要因素。结果表明，学习成绩好，家庭关系和睦，人际关系融洽，每天上网时间少，学生的网瘾程度就低。
关键词：网络成瘾 累积logistic模型 逐步回归
Abstract
This paper studies the elements that influence the net-addiction among university students. Before the data collection of the survey, the author has determined the goal of the survey is to find the factors of the net-addiction among university students, that is also to find independent variable. The author selects 16 factors as the independent variable. When setting the independent variable, taking the accuracy into consideration, the author has subdivided the net-addition into serious, general, slight and no. And cluster analysis of the hierarchical clustering method to demonstrate the reasonableness of the dependent variable values .The author has determined the main factor of the net-addition by logistic regression analysis. Through the date analysis, the author has confirmed that the academic performance, family and interpersonal relationship are the main factors. The analyse result shows that if one's academic performance is better, family and interpersonal relationship are more harmonious, and one play less internet, then one's net addition is lower.
