Skip to content

sowe9385/microLearner

 
 

Repository files navigation

Machine Learning Package for Microbiome

Build Status

Dependencies

  1. scikit-learn
  2. click
  3. pandas

Running the Commands

Bash completion

Empowered by click, microLearner supports bash auto-completion for command, subcommands, arguments and options. Users just need to run:

source bash_completion.sh

Or to make it permenant, add the above line to your ~/.bashrc file.

The bash_completion.sh file comes with the package. But you can also create or update it with the following command:

_MICROLEARNER_COMPLETE=source microLearner > bash_completion.sh

Other tips

  1. microLearner supports shorthands for subcommands as long as the shorthand is unambiguous. For example, microLearner prep --help and microLearner preprocess --help are exactly the same.

  2. microLearner is insensitive to the letter cases. For example, microLearner prep --help and MicroLearner Prep --HELP are exactly the same.

Preprocess

preprocess is the subcommand of microLearner. MinMaxScaler is the argument to the subcommand. --range, -o and -i are the options. The options for the command is insensitive to the letter cases.

microLearner preprocess MinMaxScaler --range 0 1 -i microLearner/data/otus.txt -o /tmp/foo.txt
microLearner preprocess MinMaxScaler --Range 0 1 -i microLearner/data/otus.txt -o /tmp/foo.txt

Classification

microLearner classify DecisionTreeClassifier --criterion entropy --max_features None --max_depth None --min_samples_split 2 --min_samples_leaf 1 --random_state None
microLearner classify DecisionTreeClassifier --criterion entropy --max_features None --max_depth None --min_samples_split 2 --min_samples_leaf 1 --random_state None
microLearner classify RandomForestClassifier --n_estimators 500
microLearner classify RandomForestClassifier 

Regression

microLearner regress

Clustering

microLearner cluster

About

Machine Learning Package for Microbiome

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published