MLT is a header-only Machine Learning library. It's main components are Models, which perform the regression/classification/clustering/transformation task at hand.
- Regressors : Used to learn a regression task of a continuous value (or vector of values)
- Classifiers : Used to learn a classification task on a defined set of categories, they return only the label of one class. Some also can return a score for each class.
- Clusters : Used to label and group unlabelled data
- Transformers : Learns a transformation from the data (e.g. normalization) which can be latter applied to other Model
- Samples are expected to be passed as matrices with each sample as a column, that is: [n_features, n_samples]. The same goes for the target values in supervised learning: [n_output, n_samples], and also for the predictions of the Models.