# pathClass v0.9.4

0

0th

Percentile

## Classification using biological pathways as prior knowledge

pathClass is a collection of classification methods that use information about feature connectivity in a biological network as an additional source of information. This additional knowledge is incorporated into the classification a priori. Several authors have shown that this approach significantly increases the classification performance.

## Functions in pathClass

 Name Description extractFeatures Extracts features which have been choosen by the classifier(s). fit.rrfe Reweighted Recursive Feature Elimination (RRFE) predict.graphSVM Predict Method for Graph-SVM Fits matchMatrices Matches the expression data to the adjacency matrix using the provided mapping. predict.rrfe Predict Method for RRFE Fits fit.rfe Recursive Feature Elimination (RFE) summarizeProbes Summarize probe sets predict.rfe Predict Method for RFE Fits desummarize.ranks Desummarize GeneRanks back to the corresponding probesets predict.networkBasedSVM Predict Method for Network-based SVM Fits calc.diffusionKernel Calculation of diffusion kernel matrix mapping A mapping of Refseq Protein IDs to probe set IDs for the gene expression data fit.graph.svm Implementation of a supervised classification framework introduced by Franck Rapaport et al., 2007. pathClass-package Classification with SMVs and prior knowledge fit.networkBasedSVM Implementation of the network-based Support Vector Machine introduced by Yanni Zhu et al., 2009. plot.pathClassResult Prints the result of one or more cross-validation run(s) adjacency.matrix An adjacency matrix of a random graph crossval Performs cross-validation with a specified algorithm as.adjacencyList Uses a adjacency matrix to create a adjacency list y Example class labels for the gene expression data read.hprd Parse the HPRD flat file x Example gene expression data getGeneRanks Calculate GeneRanks as used by RRFE No Results!