You are here: Symbol Reference > Classifier Namespace > Classes > TClassifier Class > TClassifier Methods > TClassifier.PrePrune Method
Data Miner
TClassifier.PrePrune Method

Pruning disables some attributes, to improve classificaiton accuracy.

function PrePrune(const b: array of TIndexRecord): double;

Disabling some attributes attributes helps reducing noise in the response of the classes. The noise usually raises with the number of different attributes. All attributes are first disabled and then enabled one by one, to check, if this improves the classification accuracy. This process is called pre-prunning. The method calls the ClassifyTest method, if internalPruning is True. Otherwise, the classification results must be returned via the OnClassifyTest event.

Copyright (c) 1999-2017 by Dew Research. All rights reserved.
What do you think about this topic? Send feedback!