CostSensitiveClassifier

A meta classifier that makes its base classifier cost-sensitive. Two methods can be used to introduce cost-sensitivity: reweighting training instances according to the total cost assigned to each class; or predicting the class with minimum expected misclassification cost (rather than the most likely class). Performance can often be improved by using a Bagged classifier to improve the probability estimates of the base classifier.

Since the classifier //normalizes// the cost matrix before applying it, it [|makes it hard] coming up with a cost matrix, e.g., to balance out imbalanced data. Here's an example: code -3 1  1  1 -6  1  0  0  0 code code 0 7 1 4 0 1 3 6 0 code But still, according to [|Weka users], one can set //the cost as to equalize the class distributions (i.e.// //achieving a 1:1 class distribution afterward)//. But this could be limited to 2-class problems.
 * input cost matrix
 * normalized cost matrix

The application of the cost matrix in MetaCost is more intuitive.

=See also=
 * MetaCost
 * CostMatrix