| Type: | Package |
| Title: | Machine Learning Assisted Network Inference |
| Version: | 1.0 |
| Author: | Mehrab Ghanat Bari |
| Maintainer: | Mehrab Ghanat Bari <m.ghanatbari@gmail.com> |
| Description: | Find dark genes. These genes are often disregarded due to no detected mutation or differential expression, but are important in coordinating the functionality in cancer networks. |
| License: | GPL-3 |
| LazyData: | TRUE |
| Depends: | e1071, stats |
| RoxygenNote: | 5.0.1 |
| NeedsCompilation: | no |
| Packaged: | 2016-09-26 17:54:07 UTC; m157258 |
| Repository: | CRAN |
| Date/Publication: | 2016-09-26 23:44:53 |
G SVM models.
Description
Returns accuracy performance of all genes. G support vector machine (SVM) classifiers trained using G different data matrixes, are used to predict labels in test data. Models are ranked based on prediction performances.
Usage
Gsvmod(dat.train, lab.train, dat.test, lab.test)
Arguments
dat.train |
Train data with G features and (k-1)*S/k samples. Parameter k comes from cross-validation scheme and is specified by user (default is 2). |
lab.train |
Class labels for train data. |
dat.test |
Test data with G features and S/k samples. |
lab.test |
Class labels for test data. |
Value
Accuracy scores for models. Each model represents one gene.
A matrix of expression values.
Description
A numeric matrix 100*20.
Usage
dat
Format
matrix.
A vector of class labels for dat.
Description
Vector length of 20.
Usage
grp
Format
vector
Select initial gene list from original data matrix.
Description
Train G-1 SVM models in k-fold cross validation scheme to select initial genes list.
Usage
intGenes(dat, grp, nfolds.out = 2, top.per = 0.05)
Arguments
dat |
Original gene expression data matrix with G rows (number of genes) and S column (number of samples). |
grp |
Class labels. |
nfolds.out |
Outer cross validation number (default is 2). |
top.per |
All genes are ranked based on their models performance and |
Value
Selected initial genes.
Examples
data(malanidata)
int <- intGenes(dat,grp)
print(int$top.genes)
Dataset for malani package
Description
A numeric matrix G*S contains gene expressions data. G are the genes (rows) and S are the samples (columns).
Usage
malanidata
Format
A matrix of numeric values, 100 genes, 20 samples and class labels.
Examples
data(malanidata)
Find best performing pairs
Description
Combine each gene in initial set with all genes in the original set. Top npair pairs are selected to construct the Q matrix.
Usage
pairmod(X, LX, theta, npair = 10)
Arguments
X |
Original gene expression data matrix. With G rows (number of genes) and S column (number of samples). |
LX |
Class labels. |
theta |
Initial gene set. |
npair |
Given a gene in initial set, top |
Value
Best (npair*G/20) performing pairs.