Case study for AutoML (CAML). Performs data cleaning and a variety of feature transformations for machine learning case studies. Uses the AutoML software TPOT, based in Scikit-Learn, for machine learning. Includes specialized options for spectral input data.
Once the source files are downloaded, the following commands can be used to download the required packages.
$ cd CAML
$ conda install numpy scipy scikit-learn pandas joblib pytorch
$ conda install -c anaconda pyyaml
$ pip install tpot
Once installed, from the home directory, the code can be run with:
$ caml.py --help
First unzip the csv file and ensure the path in the simple_input.yaml file matches where you want to store it. An example can be run by using simple-input.yaml as the request file
$ caml.py simple-input.yaml
Results will be output to a folder marked request with a datetime ID number.
CAML is released under the MIT license; see LICENSE for details.