An OptimizingLinearVectorClassifierTrainer pre-tuned with default arguments well-suited to batch training, operating on all the gradients of the Examples together.
A record of the result of applying a Classifier to a variable.
Performs iid prediction of a DiscreteVar.
Calculate the information gain between all features of Instances and the Instances' labels.
A VectorClassifier in which the score for each class is a dot-product between the observed feature vector and a vector of parameters.
An object that can create and train a LinearVectorClassifier (or train a pre-existing LinearVectorClassifier) given labeled training data.
Creates a trained naive Bayes classifier by counting feature occurrences, smoothed with pseudo-counts (m-Estimates).
An OptimizingLinearVectorClassifierTrainer pre-tuned with default arguments well-suited to online training, operating on the gradient of one Example at a time.
A LinearVectorClassifierTrainer that uses the cc.
Calculate the information gain between the binary variable "in class" / "not in class" and the binary variable "has feature" / "does not have feature" for every (label,feature) combination.
Calculate the weighted log-odds ratio: p(w|c) * log(p(w|c)/p(w|!c)) for each word w and label c.
An OptimizingLinearVectorClassifierTrainer pre-tuned with default arguments well-suited to training an L2-regularized linear SVM.
A collection of Classification results, along with methods for calculating several evaluation measures.
A Classifier in which the "input, observed" object to be classified is a VectorVar (with value Tensor1).
An object that can create and train a VectorClassifier given labeled training data.