read_metaml_results_binary: Read MetaML binary classification output

In g-antonello/gautils2: Wrapper of my functions to wrangle mostly microbiome data

Read MetaML binary classification output

Description

So far it works only with binary classification using Random Forest, and it assumes the output to have top 25 features too.

Usage

read_metaml_results_binary(output_dir, output_prefix)

Arguments

output_dir	A character vector to locate the output directory (see the code you used to launch the MetaML algorithm)
output_prefix	A character vector to know which name prefix you gave to your files

Value

A list containing all statistics divided into "full" and "top25" dataset

Examples

##### say you run something like the following code:

# python metaml/classification_thomas-manghi.py
#    input_data/data_ready.txt # input file with all variables needed in rows
#   output/Exp1 # output prefix
#   -l rf # type of learning (my function was tested only on rf)
#   -z s__ # prefix of microbiome- related variables
#   -mf 0.1 # number of ...?
#   -nt 1000 # number of trees
#    -nsl 10 # number of samples per tree leaf
#   --define 1:varible_of_interest:category_of_interest # the rest becomes 0
#   -c entropy # way to estimate the variable importance
#   -p 10 # number of folds for cross-validation
#   -r 10 # number of total runs. NB: each run is independent, while folds within a run are not independent
#    -cc 25 # re-do the analysis with this number of top features, append results on the same files

##### when this is done, go back to R and do:

# my_ML_results <- read_metaml_results_binary("output/", "Exp1")

g-antonello/gautils2 documentation built on Nov. 28, 2022, 9:39 a.m.