R中的混合效应逻辑回归模型

时间:2018-05-18 00:15:31

标签: r logistic-regression mixed-models

我正在尝试使用以下代码制作此混合效果模型:

f = read.csv2( file = "C:/Users/Customer/Desktop/R/language.csv", sep="," )

data = f[,c("wals_code", "X81A.Order.of.Subject..Object.and.Verb", 
"X87A.Order.of.Adjective.and.Noun", "family")]

data[data==""]<-NA

data<-data[complete.cases(data),]

data <- data[!data$X87A.Order.of.Adjective.and.Noun == "3 No dominant 
order", ]

data <- data[!data$X81A.Order.of.Subject..Object.and.Verb == "7 No dominant 
order", ]

View(data)

m <- glmer(X87A.Order.of.Adjective.and.Noun ~ 
X81A.Order.of.Subject..Object.and.Verb + family + (1|wals_code), data = 
data, family = binomial)

但是我收到以下错误:

警告讯息:

1:In(函数(fn,par,lower = rep.int(-Inf,n),upper = rep.int(Inf,:   未能在10000次评估中收敛

2:在checkConv(attr(opt,“derivs”)中,选择$ par,ctrl = control $ checkConv,:   无法评估缩放梯度

3:在checkConv(attr(opt,“derivs”)中,选择$ par,ctrl = control $ checkConv,:   模型未能收敛:使用4个负特征值退化Hessian

0 个答案:

没有答案