Reliability of Training Data Sets for ML Classifiers: A Lesson Learned from Mechanical Engineering.
Radmila JuricNatallia DanilchankaMehdi Gebreil MousaviPublished in: HICSS (2020)
Keyphrases
- mechanical engineering
- training set
- data sets
- training examples
- training data
- test set
- training samples
- training stage
- maximum likelihood
- decision trees
- training process
- multi layer perceptron
- class labels
- discriminative classifiers
- feature selection algorithms
- supervised learning
- support vector
- labeled training data
- linear svm
- feature set
- semi supervised learning
- classifier training
- supervised training
- training set size
- classification algorithm
- learned models
- publicly available data sets
- trained classifiers
- boosted classifiers
- benchmark data sets
- feature selection
- test data
- artificial neural networks
- classification accuracy
- learning stage
- neural network
- data streams
- feature space
- classification performances
- active learning
- linear classifiers
- unseen data
- multi class
- em algorithm
- naive bayes
- bayesian network classifiers
- training dataset