A geometric proof of the equivalence between AUC_ROC and Gini index area metrics for binary classifier performance assessment.
Paulo J. L. AdeodatoSílvio B. MeloPublished in: IJCNN (2022)
Keyphrases
- roc curve
- receiver operating characteristic
- gini index
- decision trees
- feature selection
- roc analysis
- bayes optimal
- information gain
- probability estimation
- training set
- misclassification error
- evaluation measures
- learning algorithm
- binary classifiers
- support vector
- training data
- error rate
- logistic regression
- misclassification costs
- training examples
- probability estimates
- class distribution
- classification accuracy
- multi class
- machine learning
- feature space
- feature ranking
- naive bayes
- support vector machine
- training samples
- feature set
- mutual information
- class probabilities
- svm classifier
- class labels
- text categorization