Lab 5 - Model Building Competition

Your First Data Analysis

library(caret)

# 10-fold cross-validation
train_control <- trainControl(method = "cv", number = 10)

cv_model <- train(home_valueE ~ median_incomeE,
                  data = challenge_data,
                  method = "lm",
                  trControl = train_control)

cv_model$results

  intercept     RMSE  Rsquared      MAE   RMSESD RsquaredSD    MAESD
1      TRUE 23164.79 0.8450294 17813.94 9260.733  0.1289697 6239.799

cv_model <- train(home_valueE ~ median_rentE,
                  data = challenge_data,
                  method = "lm",
                  trControl = train_control)

cv_model$results

  intercept     RMSE  Rsquared      MAE   RMSESD RsquaredSD    MAESD
1      TRUE 27115.71 0.8127802 22232.92 9173.039  0.1277109 7321.037

cv_model <- train(home_valueE ~ log(percent_collegeE),
                  data = challenge_data,
                  method = "lm",
                  trControl = train_control)

cv_model$results

  intercept     RMSE  Rsquared      MAE  RMSESD RsquaredSD    MAESD
1      TRUE 53083.21 0.5353948 43827.74 17168.8   0.262354 14650.44

cv_model <- train(home_valueE ~ median_incomeE + median_rentE + log(percent_collegeE),
                  data = challenge_data,
                  method = "lm",
                  trControl = train_control)
cv_model$results

  intercept     RMSE  Rsquared      MAE   RMSESD RsquaredSD    MAESD
1      TRUE 20299.88 0.8934533 15887.33 6976.052 0.08245509 6099.722