Partial AUC Optimisation Using Recurrent Neural Networks for Music Detection with Limited Training Data.
Pablo GimenoVictoria MingoteAlfonso Ortega GiménezAntonio MiguelEduardo LleidaPublished in: INTERSPEECH (2020)
Keyphrases
- recurrent neural networks
- training data
- classification accuracy
- neural network
- decision trees
- training set
- feed forward
- learning algorithm
- recurrent networks
- echo state networks
- reservoir computing
- naive bayes
- genetic algorithm
- artificial neural networks
- class distribution
- neural model
- feedforward neural networks
- data sets
- cascade correlation
- logistic regression
- false positives
- object detection
- prior knowledge
- test set
- complex valued
- generalization error
- nonlinear dynamic systems
- probability estimation
- test data
- detection algorithm
- receiver operating characteristic
- feature space
- feature selection
- model selection
- supervised learning