Now, let's assess the performance of the model on train and test datasets. The AUC on the train data is 0.978 and on the test data is 0.982:

# Train accuracy (AUC)
> train_pred <- predict(model.nn,occupancy_train.x)
> train_yhat <- max.col(t(train_pred))-1
> roc_obj <- pROC::roc(c(occupancy_train.y), c(train_yhat))
> pROC::auc(roc_obj)
Area under the curve: 0.9786

#Test accuracy (AUC)
> test_pred <- predict(nnmodel,occupancy_test.x)
> test_yhat <- max.col(t(test_pred))-1
> roc_obj <- pROC::roc(c(occupancy_test.y), c(test_yhat))
> pROC::auc(roc_obj)
Area under the curve: 0.9824