This presentation is about about confusion matrices, why they are confusing, and how
we can take some of that confusion away to get a better understanding of
classifier behaviour and performance, both for binary classifiers—which make
decisions where there are only two choices—and for multinomial classifiers,
where there are more than two classes to choose from.
To read more about these ideas, please see our paper
- Lovell, David, Bridget McCarron, Brendan Langfield, Khoa Tran, and Andrew P. Bradley. 2021. ‘Taking the Confusion Out of Multinomial Confusion Matrices and Imbalanced Classes’. In Data Mining, edited by Yue Xu, Rosalind Wang, Anton Lord, Yee Ling Boo, Richi Nayak, Yanchang Zhao, and Graham Williams, 16–30. Communications in Computer and Information Science. Singapore: Springer. https://doi.org/10.1007/978-981-16-8531-6_2.