在lavaan中,我有此警告消息-协方差矩阵不是正定
我正在使用lavaan来运行我的模型,并且已经收到适合我的警告消息,即协方差矩阵不是正定值。鉴于我没有统计学方面的良好背景,我想知道是否有人可以告诉我这对我的模型的影响。我是否需要删除一些潜在变量?
library(lavaan)
data <- data.frame(read.csv("Data_analysis.csv", sep=";", header=TRUE))
attach(data)
model<-" bc=~BC1+BC2+BC3+BC4+BC5+BC6 ,sc=~SC1+SC2+SC3
sust =~Sust1+Sust2,fc =~FC1+FC2+FC3+FC4+FC5 ,vis=~Visibility1+Visibility2+Visibility3+Visibility4+Visibility5+Visibility6+Visibility7,dv=~Diversity1+DIversity2+Diversity3+Diversity4+Diversity5+Diversity6, info=~Info1+Info2+Info3+Info4+Info5+Info6+Info7, so=~SO1+SO2+SO3+SO4+SO5
"
fit<-cfa(model=model, data=data)
So this was the warning message that I got.
Warning messages:
1: In lav_object_post_check(object) :
lavaan WARNING: some estimated ov variances are negative
2: In lav_object_post_check(object) :
lavaan WARNING: covariance matrix of latent variables
is not positive definite;
use lavInspect(fit, "cov.lv") to investigate.
运行我的拟合并查看模型拟合统计信息的值,我认为该值太大,但是我真的不知道如何解释该值和自由度。估计值为ML,模型拟合检验统计值为1716.008,自由度为751。这可以吗?