Random Forests
Dr. D’Agostino McGowan
1 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
The heart disease example
Recall that we are predicting whether a patient has heart disease from 13 predictors
2 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
1. Randomly divide the data in half, 149 training observations, 148 testing
set.seed(77)heart_split <- initial_split(heart, prop = 0.5)heart_train <- training(heart_split)3 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
2. Create model specification
model_spec <- rand_forest( mode = "classification", mtry = ---) %>% set_engine("ranger")mtry here is m. If we are doing bagging what do you think we set this to?
4 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
2. Create bagging specification
bagging_spec <- rand_forest( mode = "classification", mtry = 13) %>% set_engine("ranger")5 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
2. Create bagging specification
bagging_spec <- rand_forest( mode = "classification", mtry = 13) %>% set_engine("ranger")What would we change mtry to if we are doing a random forest?
5 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
2. Create Random Forest specification
rf_spec <- rand_forest( mode = "classification", mtry = 3) %>% set_engine("ranger")6 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
2. Create Random Forest specification
rf_spec <- rand_forest( mode = "classification", mtry = 3) %>% set_engine("ranger")- The default for
rand_forestisfloor(sqrt(# predictors))(so 3 in this case)
6 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
3. Fit the model
model <- fit(rf_spec, HD ~ Age + Sex + ChestPain + RestBP + Chol + Fbs + RestECG + MaxHR + ExAng + Oldpeak + Slope + Ca + Thal, data = heart_train)7 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
4. Examine how it looks in the test data
heart_test <- testing(heart_split)model %>% predict(new_data = heart_test) %>% bind_cols(heart_test) %>% conf_mat(truth = HD, estimate = .pred_class) %>% autoplot(type = "heatmap")
8 / 9
Dr. Lucy D'Agostino McGowan adapted from slides by Hastie & Tibshirani
10:00
AE 05 - Random Forests
- Open your application exercise (05) from last class
- Refit your model as a random forest
- Knit, Commit, Push. Be sure to have the final results pushed to GitHub by April 10 at noon
9 / 9