| PRTools contents |
FEATSELO
W = featselo(A,CRIT,K,T,FID)
| Input | |
| A | input dataset |
| CRIT | string name of the criterion or untrained mapping (optional, def= 'NN' 1-Nearest Neighbor error) |
| K | numner of features to select (optional, def: K=2) |
| T | validation set (optional) |
| N | Number of cross-validations (optional) |
| FID | File ID to write progress to (default [], see PRPROGRESS) |
| Output | |
| W | output feature selection mapping |
Backward selection of K features by baktracking using the branch and bound procedure on the data set A. CRIT sets the criterion used by the feature evaluation routine FEATEVAL. If the data set T is given, it is used as test set for FEATEVAL. Alternatively a number of cross-validations N may be supplied. The resulting W can be used for the selecting features of a dataset B by B*W. The selected features are stored in W.DATA and can be found by +W.
This procedure finds the optimum feature set if a monotoneous criterion is used. The use of a testset does not guarantee that.
P. M. Narendra and K. Fukunaga A Branch and Bound Algorithm for Feature Subset Selection, IEEE Trans. Computer, 26(9), pp. 917-922, September 1977
mappings, datasets, feateval, featself, featselb, featseli,
| FEATSEL, | FEATSELP, FEATSELM, PRPROGRESS |
| PRTools contents |