discriminantFuzzyPattern: Discriminant Fuzzy Pattern to filter genes
In DFP: Gene Selection

Description Usage Arguments Details Value Author(s) References Examples

discriminantFuzzyPattern discovers significant genes based on the construction of Fuzzy Patterns (FPs). The Fuzzy Patterns are built by means of applying 3 Membership Functions to the gene expression values in the matrix rmadataset.

1	discriminantFuzzyPattern(rmadataset, skipFactor = 3, zeta = 0.5, overlapping = 2, piVal = 0.9)

`rmadataset`	`ExpressionSet` with numeric values containing gene expression values (rows) of samples belonging to different categories (columns). The `ExpressionSet` also contains an `AnnotatedDataFrame` with metadata regarding the classes to which each sample belongs.
`skipFactor`	Numeric value to omit odd values (a way of normalization). Higher values imply that less samples of a gene are considered as odd. If `skipFactor=0` do NOT skip. `Default value = 3`. `Range[0,)`.
`zeta`	Threshold value which controls the activation of a linguistic label ('Low', 'Medium' or 'High'). The lower, the less posibilities of having genes with more than one assigned linguistic label. `Default value = 0.5`. `Range[0,1]`.
`overlapping`	Modifies the number of membership functions used in the discretization process. Possible values: ‘Low’, ‘Medium’, ‘High’. ‘Low’, ‘Low-Medium’, ‘Medium’, ‘Medium-High’, ‘High’. ‘Low’, ‘Low-Medium’, ‘Low-Medium-High’, ‘Medium’, ‘Medium-High’, ‘High’. `Default value = 2`.
`piVal`	Controls the degree of exigency for selecting a gene as a member of a Fuzzy Pattern. `Default value = 0.9`. `Range[0,1]`.

The discriminantFuzzyPattern function works in a 4-step process:

Calculates the Membership Functions. These functions are used in the next step to discretize gene expression data.
Discretizes the gene expression data (float values) into ‘Low’, ‘Medium’ or ‘High’ labels.
Calculates a Fuzzy Pattern for each category. To do this, a given percentage of the samples belonging to a category must have the same label (‘Low’, ‘Medium’ or ‘High’).
Calculates the Discriminant Fuzzy Pattern (DFP) that includes those genes present in two or more FPs with different assigned labels.

`membership.functions`	Membership functions to determine the discret value corresponding to a given gene expression level.
`discrete.values`	Discrete values according to the overlapping parameter after discretizing the gene expression values. Includes an attribute `types` which determines the category of each sample.
`fuzzy.patterns`	Genes belonging to each Fuzzy Patterns. There are one FP for each class. Includes an attribute `ifs` with the Impact Factor for each category.
`discriminant.fuzzy.pattern`	Genes belonging to the final DFP. Includes an attribute `ifs` with the Impact Factor for each category.
`params`	The parameters used to tune the algorithm (as arguments in the function).

Rodrigo Alvarez-Gonzalez
Daniel Glez-Pena
Fernando Diaz
Florentino Fdez-Riverola
Maintainer: Rodrigo Alvarez-Gonzalez <rodrigo.djv@uvigo.es>

F. Diaz; F. Fdez-Riverola; D. Glez-Pena; J.M. Corchado. Using Fuzzy Patterns for Gene Selection and Data Reduction on Microarray Data. 7th International Conference on Intelligent Data Engineering and Automated Learning: IDEAL 2006, (2006) pp. 1095-1102

#########################################
############ Get sample data ############
#########################################
library(DFP)
data(rmadataset)

#########################################
# Filters the most representative genes #
#########################################
res <- discriminantFuzzyPattern(rmadataset)
summary(res)