Skip to content
Synthetic Minority Over-Sampling Technique for Regression | PyPI
Python
Branch: master
Clone or download
Latest commit 20c82e0 Dec 22, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data
examples Update, version 0.1.0 Dec 23, 2019
media Add README banner 0.0.9 Dec 10, 2019
smogn Input check fix, version 0.1.0 Dec 23, 2019
.gitignore Update .gitignore Nov 19, 2019
.travis.yml Update .travis.yml Dec 19, 2019
LICENSE Add example data Nov 19, 2019
MANIFEST.in Update MANIFEST, include all examples 0.0.9 Dec 10, 2019
README.md Mention requirements Dec 23, 2019
UPDATES Update UPDATES, version 0.1.0 Dec 23, 2019
setup.py Update, version 0.1.0 Dec 23, 2019

README.md

Synthetic Minority Over-Sampling Technique for Regression with Gaussian Noise

PyPI version License: GPL v3 Build Status Codacy Badge GitHub last commit

Description

A Python implementation of Synthetic Minority Over-Sampling Technique for Regression with Gaussian Noise (SMOGN). Conducts the Synthetic Minority Over-Sampling Technique for Regression (SMOTER) with traditional interpolation, as well as with the introduction of Gaussian Noise (SMOTER-GN). Selects between the two over-sampling techniques by the KNN distances underlying a given observation. If the distance is close enough, SMOTER is applied. If too far away, SMOTER-GN is applied. Useful for prediction problems where regression is applicable, but the values in the interest of predicting are rare or uncommon. This can also serve as a useful alternative to log transforming a skewed response variable, especially if generating synthetic data is also of interest.

Features

  1. The only open-source Python supported version of Synthetic Minority Over-Sampling Technique for Regression.

  2. Supports Pandas DataFrame inputs containing mixed data types, auto distance metric selection by data type, and optional auto removal of missing values.

  3. Flexible inputs available to control the areas of interest within a continuous response variable and friendly parameters for over-sampling synthetic data.

  4. Purely Pythonic, developed for consistency, maintainability, and future improvement, no foreign function calls to C or Fortran, as contained in original R implementation.

Requirements

  1. Python 3
  2. NumPy
  3. Pandas

Installation

## install pypi release
pip install smogn

## install developer version
pip install git+https://github.com/nickkunz/smogn.git

Usage

## load libraries
import smogn
import pandas

## load data
housing = pandas.read_csv(
    
    ## http://jse.amstat.org/v19n3/decock.pdf
    "https://raw.githubusercontent.com/nickkunz/smogn/master/data/housing.csv"
)

## conduct smogn
housing_smogn = smogn.smoter(
    
    data = housing, 
    y = "SalePrice"
)

Examples

  1. Beginner
  2. Intermediate
  3. Advanced

License

© Nick Kunz, 2019. Licensed under the General Public License v3.0 (GPLv3).

Contributions

SMOGN is open for improvements and maintenance. Your help is valued to make the package better for everyone.

Reference

Branco, P., Torgo, L., Ribeiro, R. (2017). SMOGN: A Pre-Processing Approach for Imbalanced Regression. Proceedings of Machine Learning Research, 74:36-50. http://proceedings.mlr.press/v74/branco17a/branco17a.pdf.

You can’t perform that action at this time.