Predicting the embedding coordinates via the trained SVM model.

Source: R/projection.R

Predicting the embedding coordinates via the trained SVM model.

ProjectNewdata(
  feature.mat,
  model,
  do.norm = "L2",
  int.fun = stats::median,
  cores = -1
)

Arguments

feature.mat

A data.frame containing signature scores, rows are cells, columns are features.

model

The trained learners returned from `FitEnsembleSVM()`.

do.norm

Whether normalize the feature matrix. L1, L2, NULL. Default: 'L2'

int.fun

The function for integration of predicted embedding by different learners: mean, median, or any self defined function given a vector and returns a value. Default: median.

cores

number of threads for prediction, -1 means all available threads. Default: -1

Value

CellProject object containing source data and predicted embedding matrix.