Performaning feature selection in forward stepwise selection maner.
Project description
ForwardStepwiseFeatureSelection aims to select the best features or the subset of features in machine learning tasks according to corresponding score with other incredible packages like numpy, pandas and sklearn.
This package is inspired by: PyData DC 2016 | A Practical Guide to Dimensionality Reduction Vishal Patel October 8, 2016
- Examples: https://github.com/HindyDS/ForwardStepwiseFeatureSelection/tree/main/examples
- Email: hindy888@hotmail.com
- Source code: https://github.com/HindyDS/ForwardStepwiseFeatureSelection/tree/main/ForwardStepwiseFeatureSelection
- Bug reports: https://github.com/HindyDS/ForwardStepwiseFeatureSelection/issues
It requires at least six arguments to run:
- estimators: machine learning model
- X (array): features space
- y (array): target
- cv (int): number of folds in a (Stratified)KFold
- scoring (str): see https://scikit-learn.org/stable/modules/model_evaluation.html
Optional arguments:
- max_trial (int): number of trials that you wanted RFS to stop searching
- tolerance (int): how many times RFS can fail to find better subset of features
- least_gain (int): threshold of scoring metrics gain in fraction
- max_feats (int): maximum number of features
- prior (list): starting point for RFS to search, must be corresponds to the columns of X
- exclusions (nested list): if the new selected feature is in one of the particular subpool (list in the nested list), then the features in that particular subpool with no longer be avalible to form any new subset in the following trials
- n_jobs (int): Number of jobs to run in parallel.
- n_digit (int): Decimal places for scoring
- verbose (int): Level of verbosity of RFS
If you have any ideas for this packge please don't hesitate to bring forward!
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Close
Hashes for ForwardStepwiseFeatureSelection-1.3.9.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 880b7d59987c84445af2ce0317d606aba24202023996207c31bc15b51cd957bd |
|
MD5 | 5f7a57f8219a4a9a1185db1471cfe304 |
|
BLAKE2b-256 | 87bdb00876e09e01c8154c531ed12c8cabea0b923896b2269983d92fdd321081 |
Close
Hashes for ForwardStepwiseFeatureSelection-1.3.9-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | fce03d2d65204c71d50b1d540ba1341f08b0a69a5c694c45724974ef020908f3 |
|
MD5 | 729b19a8c41dd38ca03749e74e38e0db |
|
BLAKE2b-256 | 8204535dcf29e6319436573783b1f05acca603c631258f8843c8623967f7ddad |