aikit

Automatic Tool Kit for Machine Learning and Datascience.

The objective is to provide tools to ease the repetitive part of the DataScientist job and so that he/she can focus on modelization. This package is still in alpha and more features will be added. Its mains features are:

• improved and new "scikit-learn like" transformers ;
• GraphPipeline : an extension of sklearn Pipeline that handles more generic chains of tranformations ;
• an AutoML to automatically search throught several transformers and models.

Full documentation is available here: https://aikit.readthedocs.io/en/latest/

GraphPipeline

The GraphPipeline object is an extension of sklearn.pipeline.Pipeline but the transformers/models can be chained with any directed graph.

The objects takes as input two arguments:

• models: dictionary of models (each key is the name of a given node, and each corresponding value is the transformer corresponding to that node)
• edges: list of tuples that links the nodes to each other

Example:

gpipeline = GraphPipeline(
    models = {
        "vect": CountVectorizerWrapper(analyzer="char",
                                       ngram_range=(1, 4),
                                       columns_to_use=["text1", "text2"]),
        "cat": NumericalEncoder(columns_to_use=["cat1", "cat2"]), 
        "rf": RandomForestClassifier(n_estimators=100)
    },
    edges = [("vect", "rf"), ("cat", "rf")]
)

AutoML

Aikit contains an AutoML part which will test several models and transformers for a given dataset.

For example, you can create the following python script run_automl_titanic.py:

from aikit.datasets import load_dataset, DatasetEnum
from aikit.ml_machine import MlMachineLauncher

def loader():
    dfX, y, *_ = load_dataset(DatasetEnum.titanic)
    return dfX, y

def set_configs(launcher):
    """ modify that function to change launcher configuration """
    launcher.job_config.score_base_line = 0.75
    launcher.job_config.allow_approx_cv = True
    return launcher

if __name__ == "__main__":
    launcher = MlMachineLauncher(base_folder = "~/automl/titanic", 
                                 name = "titanic",
                                 loader = loader,
                                 set_configs = set_configs)
    launcher.execute_processed_command_argument()

And then run the command:

python run_automl_titanic.py run -n 4

To run the automl using 4 workers, the results will be stored in the specified folder You can aggregate those result using:

python run_automl_titanic.py result