bestGrouping: description of function randomForest

In stela2502/BioData: The package allows to annotate a numeric data.table with col and row data

Description

The function is using a randomForest classifier with 2000 trees to classify the given data using the given grooping

All groups that fail to be prediceted using the random forest are deemed ungrouped.

All groups where less than 50 percent of the total samples geting classified as being from that group fail.

Arguments

x	the single cells ngs object
group	a vector of sample columns that should be checked (the most complex is used only)
bestColname	the column name to store the best grouping in
cutoff	the cutoff percentage where all groups showing less than this percentacge of remapped samples are dropped
ntree	for the rf process (default 2000)
...	additional variables for the randomForest call

Value

a distRF object to be analyzed by pamNew

Examples

## lets data be a source BioData object and selection be a smaller one with only selected samples 
## for which a grouping has been produced named 'new_perfect_grouping'
## Not run:
#predictor <- bestGrouping( selection, 'new_perfect_grouping' )
#data$samples$new_perfect_grouping <- predict(predictor, as.matrix(t(data$dat)))
## End(**Not run**)

stela2502/BioData documentation built on Feb. 23, 2022, 5:47 a.m.