Hi,
For a logit fit model, I divided my study population into 80% and 20%.
I used the 80% group for training and 10-fold cross validation.
For external validation of the optimized model, I used the Test set containing 20% case group and tabulated the results obtained from the Test set for model performance.
Now I want to use XGBoost with the same training and test sets.
Can I specify a fixed set of cases as a Test set in R without creating a random test set?
XGBoost has the famous mushroom dataset. Two groups are introduced here, agaricus.train and agaricus.test:
data(agaricus.train, package = 'xgboost') data(agaricus.test, package = 'xgboost') train
If you are using R: set. seed(1234) before data splitting will ensure reproducible work. The hard way is to split data and write to files for later use: i mean using several algorithms.
Use machine learning train and task and run all the machine learning models using a for loop and see the accuracy.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Fixatives