These webpages give some impressions of the usage and what the implemented algorithms can perform. Please consult the Scaladoc API for detailed and up-to-date information on the algorithms, e.g. the extensive list of implemented hyper parameters.
Individual pages describe the respective algorithms in more detail and provide examples.
Classification
The basic usage of classifiers is as follows:
/**
* @param X_train List of feature vectors for training
* @param y_train List of labels
* @param X_test List of feature vectors for prediction
*/
val clf = new Classifier() // optional hyperparameters as arguments
clf.train(X_train, y_train) // perform the training
val y_pred = clf.predict(X_test) // make a prediction
val diag = clf.diagnostics // obtain a map of metrics for algorithm training
Regression
The basic usage of regressors is as follows:
/**
* @param X_train List of feature vectors for training
* @param y_train List of labels
* @param X_test List of feature vectors for prediction
*/
val reg = new Regressor() // optional hyperparameters as arguments
reg.train(X_train, y_train) // perform the training
val y_pred = reg.predict(X_test) // make a prediction
val diag = reg.diagnostics // obtain a map of metrics for algorithm training
Clustering
Clustering or unsupervised classification algorithms are applied to data without or very little or insufficient training data. They are meant to find patterns in the data with a minimal amount of domain knowledge required to steer the algorithm.
The basic usage of clustering algorithms is as follows:
/**
* @param X_train List of feature vectors for training
* @param X_test List of feature vectors for prediction
*/
val clu = new Clustering() // optional hyperparameters as arguments
clu.train(X_train) // perform the training
val y_pred = clu.predict(X_test) // make a prediction