Given a value matrix (features are rows, samples are columns), and sample classes,
find those things that are not zero in at least a certain
number of samples in one of the classes, and keep those features for further
processing.
Usage
keep_non_zero_percentage(
data_matrix,
sample_classes = NULL,
keep_num = 0.75,
zero_value = 0,
all = FALSE
)
Arguments
- data_matrix
the matrix of values to work with
- sample_classes
the classes of each sample
- keep_num
what number of samples in each class need a non-zero value (see Details)
- zero_value
what number represents zero values
- all
is this an either / or OR does it need to be present in all?
Details
The number of samples that must be non-zero can be expressed either as a whole
number (that is greater than one), or as a fraction that will be be multiplied
by the number of samples in each class to get the lower limits for each of the classes.