mlpkg 0.3

ML package

ML Package

A Package through which multiple things can be done in a go like fitting a model of your own choice (regression/classification) and get final results in terms of model accuracy/ score/ mse/ mae/ confusion matrix and direct submission file which we need to upload in case of competitions.

Two things can be achieved from this package:

1. Results after fitting any Regression/classification algorithm of your own choice. Function to be used - train_model
2. Final submission file to be submitted directly in competitions. Function to be used - return_csv

Requirements:

1. sklearn
2. numpy
3. pandas
4. Pre-processed data to be passed - data already cleaned, splitted into train and test
5. In train_model function, parameters to be passed:
   • x_train, y_train, x_test, y_test
   • model object
   • regression=True (if addressing regression problem)
   • regression=False (if addressing classification problem (default is True))
6. In return_csv function, parameters to be passed:
   • x_train, y_train, x_test, y_test
   • model object
   • sample submission file (dataframe)
   • target column name (string)
   • file location name (filepath where to be saved)

Installation & Usage

1. Make sure that your pip version is up-to-date: pip install --upgrade pip. Check version with pip -V.
2. Select the correct package:
   • There are two packages (two versions of the package )and you should SELECT ONLY ONE OF THEM which is the latest one.
   • Install using pip install mlpkg with latest version
3. Import the package and use its functions: from ML_Utility import ML_package