splendid: SuPervised Learning ENsemble for Diagnostic IDentification

Provides a bootstrapping and ensemble framework for supervised learning analyses using multiclass classification algorithms for modelling, prediction, and evaluation. Predicted classes are evaluated under metrics such as log loss, AUC, F1-score, Matthew's correlation coefficient, and accuracy. Discrimination and reliability plots visualize the classifier performances. The .632+ estimator is implemented for the log loss error rate.

See also

Useful links:

• https://github.com/AlineTalhouk/splendid

• https://alinetalhouk.github.io/splendid/

• Report bugs at https://github.com/AlineTalhouk/splendid/issues

Author

Maintainer: Derek Chiu dchiu@bccrc.ca

Authors:

• Aline Talhouk atalhouk@bccrc.ca

• Dustin Johnson djohnson@bccrc.ca