sigFeature: Significant Feature Selection by using SVM-RFE & t-statistic.
In sigFeature: sigFeature: Significant feature selection using SVM-RFE & t-statistic

Description Usage Arguments Details Value Author(s) References See Also Examples

Significant Feature selected by using SVM-RFE and t-statistic.

1	sigFeature(X,Y)

`X`	n-by-d data matrix to train (n samples/patients, d features/genes)
`Y`	vector of class labels -1 or 1's (for n samples/patients)

The idea of "sigFeature" (Significant Feature Selection) begins from the lack of support vector machine recursive feature (SVM-RFE) method. The feature ranked by the SVM-RFE algorithm may or may not be differentially significant among the classes (in case of binary classification). Significant features which have good classification accuracy and which are differentially significant have an impotent role in biological aspect.

In data mining and optimisation, the feature selection is a very active field of research where the SVM-RFE is distinguished as one of the most effective methods. This is a greedy method that only hopes to find the best possible combination for classification. In this algorithm, greedy method similar to SVM-RFE is used. The prime intention of this algorithm is to enumerate the ranking weights for all the features and sort the features according to weight vectors as the basis for classification. The coefficient and the expression mean differences between two compared groups are used to calculate the weight value of that particular feature. The iteration process is followed by backward removal of the feature. The iteration process is continued until there is only one feature remaining in the dataset. The smallest ranking weight will be removed by the algorithm while the feature variables with significant impact remains. Finally, the feature variables will be listed in the descending order of descriptive difference degree.

returns the feature list.

Pijush Das <topijush@gmail.com>, et al.

1. Karatzoglou, Alexandros, David Meyer, and Kurt Hornik. "Support vector machines in R." (2005).

2. O'Donnell RK, Kupferman M, Wei SJ, Singhal S et al. Gene expression signature predicts lymphatic metastasis in squamous cell carcinoma of the oral cavity. Oncogene 2005 Feb 10;24(7):1244-51.

SVM, predict.penSVM

#Data set taken from GSE2280
library(SummarizedExperiment)
data(ExampleRawData, package="sigFeature")

x  <- t(assays(ExampleRawData)$counts)
y  <- colData(ExampleRawData)$sampleLabels


#Number of features are reduced to minimized the build time.
x <- x[ , 1:100]

#Feature selection with sigFeature functio.
system.time(sigfeatureRankedList <- sigFeature(x,y))
str(sigfeatureRankedList)