Map a dataset on a K-NN classifier
F = KNN_MAP(A,W)
| A|| Dataset|
| W|| K-NN classifier trained by KNNC |
| F|| Posterior probabilities|
Maps the dataset A by the K-NN classifier W on the [0,1] interval for each of the classes that W is trained on. The posterior probabilities, stored in F, are computed in the following ways
soft labeled training set: the normalised average of the soft labels of the K neighbors. crisp labeled training set, K = 1: normalisation of sigm(log(F)) with F(1:C) = sum(NN_Dist(1:C))./NN_Dist(1:C) - 1 in which C is the number of classes and NN_Dist stores the distance to the nearest neighbor of each class. crisp labeled training set, K > 1: normalisation of (N(1:C) + 1)/(K+C), in which N stores the number of objects per class within the K first neighbors.
This routine is called automatically to determine A*W if W is trained by KNNC.
Warning: Class prior probabilities in the dataset A are neglected.
mappings, datasets, knnc, testk,
|This file has been automatically generated. If badly readable, use the help-command in Matlab.|