mlpiper 1.0.0

An ML Pipeline to be executed by 'mlpiper' tool

README

This is an example of a machine-learning pipeline that is implemented by the 'mlpiper' platform.

This specific example is no more than a simple sklearn implementation, without using the built-in sklearn Pipeline structure.

The pipeline description is provided by the pipeline/pipeline.json file.

Setup

1. Browse to ml_pipeline_examples/mlpiper_example

   • > cd ml_pipeline_examples/mlpiper_example

2. Create virtual env, whose name is 'mlpiper', or anything else (If you've already created it, use workon to activate it):

   • > mkvirtualenv --python=3 mlpiper

3. Install 'mlpiper' (Skip if already installed):

   • > pip install -r requirements.txt

Run Example

(Note: Please make sure the environment was setup properly. Refer to 'Setup' section above for more information)

1. Browse to ml_pipeline_examples/mlpiper_example

   • > cd ml_pipeline_examples/mlpiper_example

2. Run the run.sh script:

   • > run.sh

Results

...
Log Loss: 0.6571923425816221
ROC AUC Score: 0.6350086940964011
...
real	0m2.199s
user	0m2.528s
sys		0m0.295s