Assignment 4.knwf (138.0 KB)
It is for the Neural Network, and I ned to figure out the code for the R part.
I need help with the RNN part. Thank you so much
Neural Networks
#install.packages(neuralnet)
require(neuralnet)
Create two classification variables
df$yes ← c(df$Y == “1”)
df$no ← c(df$Y == “0”)
Partition the data into training and validation sets
70% of the sample size
smp_size ← floor(0.70 * nrow(df))
set the seed to make the partition reproducible
set.seed(123)
train_ind ← sample(c(1:nrow(df)), size = smp_size)
train ← df[train_ind, ]
validation ← df[-train_ind, ]
Train the ANN
nn ← neuralnet(yes + no ~ X1 + X2 + X3 + X4 + X5,
train, hidden=c(5), linear.output = FALSE)
plot(nn, rep=“best”)
summary(nn)
Predict Y for the validation dataset
predict ← compute(nn, validation[,1:5])$net.result
Extract predicted values on validation set and construct a confusion matrix
net.prediction ← c(“1”,“0”)[apply(predict,1,which.max)]
predict.table ← table(validation$Y, net.prediction)
predict.table
Alternatively see more diagnostics with a confusion matrix function
#require(caret)
#require(e1071)
#confusionMatrix(predict.table)
#dim(validation)
predict.df ← as.data.frame(predict)
names(predict.df) ← c(“p_1”, “p_0”)
p_validation ← cbind(validation, predict.df, net.prediction)
knime.out ← p_validation
But we can give hints…