RNN错误“ $运算符对原子向量无效”

Question

你好，我是rnn计算的新手，我在训练网络时遇到问题。这是我的代码：

df_train <- df[df$reportingMonth<="2017-12-01",]
df_test <- df[df$reportingMonth>"2017-12-01",]

X <- array(c(df_train$DebtToAssetsRatio, df_train$taille, log(df_train$current)), dim=c(708,48,3))
Y <- matrix(df_train$insolvency, nrow=708, ncol=48)

nnet <- trainr(Y=Y, X=X, 1, learningsrate=0.1)

我收到以下消息：

retraining, all options except X, Y and the model itself are ignored, error are reseted
Error: $ operator is invalid for atomic vectors

它可能来自数组X和Y的定义，但我看不到……有人有想法吗？