Skip to main content

A Python implementation of Synthetic Minority Over-Sampling Technique for Regression with Gaussian Noise (SMOGN)

Project description

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.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

smogn-0.1.2.tar.gz (185.8 kB view details)

Uploaded Source

Built Distribution

smogn-0.1.2-py3-none-any.whl (30.3 kB view details)

Uploaded Python 3

File details

Details for the file smogn-0.1.2.tar.gz.

File metadata

  • Download URL: smogn-0.1.2.tar.gz
  • Upload date:
  • Size: 185.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/2.0.0 pkginfo/1.5.0.1 requests/2.21.0 setuptools/41.6.0 requests-toolbelt/0.9.1 tqdm/4.31.0 CPython/3.7.4

File hashes

Hashes for smogn-0.1.2.tar.gz
Algorithm Hash digest
SHA256 6555b907f2c9df223eae8813abd09054ad6491fc8509a23fccc9d578b3e76d89
MD5 796b6f3b8d9b5675b15445fa7225531c
BLAKE2b-256 b7fbdb39a4c4df49bc1754439027978e064191447b8979e69692833a6a4ccdb1

See more details on using hashes here.

File details

Details for the file smogn-0.1.2-py3-none-any.whl.

File metadata

  • Download URL: smogn-0.1.2-py3-none-any.whl
  • Upload date:
  • Size: 30.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/2.0.0 pkginfo/1.5.0.1 requests/2.21.0 setuptools/41.6.0 requests-toolbelt/0.9.1 tqdm/4.31.0 CPython/3.7.4

File hashes

Hashes for smogn-0.1.2-py3-none-any.whl
Algorithm Hash digest
SHA256 3df02540674d2164713e661446ff73994f98bb3029d5ea7ad0107ffedb700fec
MD5 654c59106ee9e47d728c028a7c8440e6
BLAKE2b-256 fea6a3f78f5e2e18fa302fd23a32019908657113faa917463a9d49cbe9d20625

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page