Feature Selection Criterion

Select a feature selection criterion to be used in optimizing the predictor set.

Available criteria for optimizing the predictor set are described below:

Method	Description
Minimum Redundancy Maximum Relevance	This option requires that features within a subset of predictors are maximally dissimilar to each other while still exceeding a specified level of mutual information with the target phenotype . See Ding and Peng (2005) 1 for more information.
Correlation Ratio	Calculates the correlation of a subset of predictors to the target phenotype and the correlation of the subsets to each other. The ratio of the correlations is used as the selection criterion.

---

1

Ding, C and Peng, H. (2005) Minimum redundancy feature selection from microarray gene expression data. J Bioinformatics and Computational Biology 3: 185-205.

To Specify the Criterion for Feature Selection:

Make sure that the Optimize the continuous predictor set check box is checked.

Select the desired option using the drop-down menu.

For additional information:

Refer to http://en.wikipedia.org/wiki/Feature_selection .