Determine hyperparameters for new classifiers.

[tsalo]

Should just be a matter of looking up each classifier for its inputs and the typical ranges used.

[tsalo]

Classifiers for which we need hyperparameters:

• Multilabel k-nearest neighbors (ML-kNN), implemented in MULAN
• RAndom k-labELsets (RAkEL), implemented in MULAN
• Deep multilabel (DeepML), implemented in MEKA
• Hierarchical multilabel classifier (HMC), implemented in MULAN
• Ensemble of Classifier Chains (ECC), implemented in MEKA
• Expectation Maximization (EM), implemented in MEKA
• Binary Relevance (BR), implemented in MEKA

[tsalo]

Hyperparameters determined and grid-searched.