stam.predict {stam} | R Documentation |
StAM analysis on new data using a given model fit.
stam.predict(fit, expression.matrix, classifications = NULL, testset = NULL)
fit |
stamFit object containing a trained model |
expression.matrix |
matrix or exprSet containing new data |
classifications |
character vector specifying class names per sample. You may either specify one class per sample in the expression.matrix, or one class per training sample (all but the testset). |
testset |
indeces of samples not yued in training |
stam.predict uses an object returned by stam.fit to perform a structured analysis of the new expression data provided. Thereby, it uses all classifiers in the leaf nodes to provide classification results in these for each sample. In addition, weighted sums in inner nodes are computed to provide classification results for the whole graph.
An object of type stamPrediction is returned. You may use print, plot and image
functions to further investigate the results. Information on node classifiers are
obtained through the print methods on elements of the nodes
slot.
Claudio Lottaz
stamPrediction-class
, plot.stamPrediction
,
image.stamPrediction
, stam.fit
,
stam.writeHTML
# load and prepare data golubTest <- golubNorm[,39:72] data(golubTrain.fit) # compute predictions golubTest.pred <- stam.predict(golubTrain.fit, golubTest, pData(golubTest)[,"ALL.AML"]) golubMerge.pred <- stam.predict(golubTrain.fit, golubNorm, pData(golubNorm)[,"ALL.AML"], testset=39:72) # further investigate print(golubTest.pred) plot(golubTest.pred) ## Not run: map <- image(golubMerge.pred, outfile="golubMerge") cat("<HTML><BODY><MAP NAME='image_map'>", map, "</MAP>", "<IMG SRC='golubMerge_pred_img.png' USEMAP='#image_map'></BODY></HTML>\n", file="pred_img.html") browseURL(paste("file://", getwd(), "/pred_img.html", sep="")) ## End(Not run)