基于变量选择的网络直播影响因素分析

Analysis of Influencing Factors in Live Webcast based on Variable Selection

作为一种新兴的社交媒体,网络直播在近年来的发展愈演愈烈。网络直播何以有如此巨大的发展规模和潜力,有哪些影响因素促进和加速了它的发展?基于子集选择法和系数压缩法的变量选择方法可以对此问题做详细的定量分析以筛选出显著变量。根据数据特征,选定运用Cp准则、AIC准则、BIC准则和LARS、AdaptiveLASSO等方法进行变量选择,通过横向比较变量选择结果得出影响网络直播火爆发展的关键因素,以对网络直播行业提供合理的产业导向信息。

As an emerging social media,webcasting has intensified in recent years.Why does webcasting have such a huge scale of development and potential,and what influence factors have contributed to and accelerated its development?The variable selection method based on the subset selection method and the coefficient compression method can perform detailed quantitative analysis on this problem to screen out significant variables.According to the data characteristics,the Cp criterion,AIC criterion,BIC criterion,LARS,Adaptive LASSO and other methods are selected for variable selection.Through the horizontal comparison of variable selection results,the key factors affecting the rapid development of webcasting are obtained,so as to provide reasonable industry-oriented information for the live broadcast industry.