Tips

library(keras)
data <- read.csv('~/Desktop/data/CTG.csv', header=T)
data$NSP <- data$NSP -1

# Turn `data` into a matrix
data <- as.matrix(data)
dimnames(data) <- NULL  

# Normalize the  data
data[,1:21] <- normalize(data[,1:21])

# Partition
set.seed(1234)
ind <- sample(2, nrow(data), replace = T, prob=c(.7, .3))
training <- data[ind==1, 1:21]
test <- data[ind==2, 1:21]
trainingtarget <- data[ind==1, 22]
testtarget <- data[ind==2, 22]

# One hot encoding 
trainLabels <- to_categorical(trainingtarget)
testLabels <- to_categorical(testtarget)

# model_one
model <- keras_model_sequential()
model %>% 
  layer_dense(units = 8, activation = 'relu', input_shape = c(21)) %>%
  layer_dense(units = 3, activation = 'softmax')  

# Compile the model 
model %>% compile(loss = 'categorical_crossentropy', 
                  optimizer = 'adam',
                  metrics = 'accuracy')

# Fit model
history <- model %>% fit(training, 
                         trainLabels, 
                         epochs = 200,  
                         batch_size = 32,
                         validation_split = 0.2)