New Ideas

Classifier

• nearest mean classifier
• Euclidean distance classifier
• naive bayes classifier
• quadratic discriminant classifier (QDA)
• regularized discriminant classifier (RDA)

Feature Selection

• Relief
• Forward Selection with Gram-Schmidt Orthogonalization

Ensamble

• Bagging
• AdaBoost

Multiclass

• one-against-rest
• one-against-one

New Abstraction Layer

• Learner (learn(x, y=None))
  • Predictor (pred(t))
    • Classifier
      • LinearClassifier (w() and bias())
    • Regressor
      • LinearRegressor (beta() and beta0())
  • Reducer (transform(t, k=None))

Clustering algorithms from Learner?

Various

• Force kernel parameters in kernel algorithms?