<p align="center">
<img src="./docs/img/mrmr_logo_white_bck.png" alt="drawing" width="450"/>
</p>
## What is mRMR
*mRMR*, which stands for "minimum Redundancy - Maximum Relevance", is a feature selection algorithm.
## Why is it unique
The peculiarity of *mRMR* is that it is a **minimal-optimal** feature selection algorithm. <br/>
This means it is designed to find the smallest relevant subset of features for a given Machine Learning task.
Selecting the minimum number of useful features is desirable for many reasons:
- memory consumption,
- time required,
- performance,
- explainability of results.
This is why a minimal-optimal method such as *mrmr* is often preferable.
On the contrary, the majority of other methods (for instance, Boruta or Positive-Feature-Importance) are classified as **all-relevant**,
since they identify all the features that have some kind of relationship with the target variable.
## When to use mRMR
Due to its efficiency, *mRMR* is ideal for practical ML applications,
where it is necessary to perform feature selection frequently and automatically,
in a relatively small amount of time.
For instance, in **2019**, **Uber** engineers published a paper describing how they implemented
*mRMR* in their marketing machine learning platform [Maximum Relevance and Minimum Redundancy Feature Selection Methods for a Marketing Machine Learning Platform](https://eng.uber.com/research/maximum-relevance-and-minimum-redundancy-feature-selection-methods-for-a-marketing-machine-learning-platform/).
## How to install this package
You can install this package in your environment via pip:
<pre>
pip install mrmr_selection
</pre>
And then import it in Python through:
<pre>
import mrmr
</pre>
## How to use this package
This package is designed to do *mRMR* selection through different tools, depending on your needs and constraints.
Currently, the following tools are supported (others will be added):
- **Pandas**
- **Polars**
- **Spark**
- **Google BigQuery**
The package has a module for each supported tool. Each module has *at least* these 2 functions:
- `mrmr_classif`, for feature selection when the target variable is categorical (binary or multiclass).
- `mrmr_regression`, for feature selection when the target variable is numeric.
Let's see some examples.
#### 1. Pandas example
You have a Pandas DataFrame (`X`) and a Series which is your target variable (`y`).
You want to select the best `K` features to make predictions on `y`.
```python
# create some pandas data
import pandas as pd
from sklearn.datasets import make_classification
X, y = make_classification(n_samples = 1000, n_features = 50, n_informative = 10, n_redundant = 40)
X = pd.DataFrame(X)
y = pd.Series(y)
# select top 10 features using mRMR
from mrmr import mrmr_classif
selected_features = mrmr_classif(X=X, y=y, K=10)
```
Note: the output of mrmr_classif is a list containing K selected features. This is a **ranking**, therefore, if you want to make a further selection, take the first elements of this list.
#### 2. Polars example
```python
# create some polars data
import polars
data = [(1.0, 1.0, 1.0, 7.0, 1.5, -2.3),
(2.0, None, 2.0, 7.0, 8.5, 6.7),
(2.0, None, 3.0, 7.0, -2.3, 4.4),
(3.0, 4.0, 3.0, 7.0, 0.0, 0.0),
(4.0, 5.0, 4.0, 7.0, 12.1, -5.2)]
columns = ["target", "some_null", "feature", "constant", "other_feature", "another_feature"]
df_polars = polars.DataFrame(data=data, schema=columns)
# select top 2 features using mRMR
import mrmr
selected_features = mrmr.polars.mrmr_regression(df=df_polars, target_column="target", K=2)
```
#### 3. Spark example
```python
# create some spark data
import pyspark
session = pyspark.sql.SparkSession(pyspark.context.SparkContext())
data = [(1.0, 1.0, 1.0, 7.0, 1.5, -2.3),
(2.0, float('NaN'), 2.0, 7.0, 8.5, 6.7),
(2.0, float('NaN'), 3.0, 7.0, -2.3, 4.4),
(3.0, 4.0, 3.0, 7.0, 0.0, 0.0),
(4.0, 5.0, 4.0, 7.0, 12.1, -5.2)]
columns = ["target", "some_null", "feature", "constant", "other_feature", "another_feature"]
df_spark = session.createDataFrame(data=data, schema=columns)
# select top 2 features using mRMR
import mrmr
selected_features = mrmr.spark.mrmr_regression(df=df_spark, target_column="target", K=2)
```
#### 4. Google BigQuery example
```python
# initialize BigQuery client
from google.cloud.bigquery import Client
bq_client = Client(credentials=your_credentials)
# select top 20 features using mRMR
import mrmr
selected_features = mrmr.bigquery.mrmr_regression(
bq_client=bq_client,
table_id='bigquery-public-data.covid19_open_data.covid19_open_data',
target_column='new_deceased',
K=20
)
```
## Reference
For an easy-going introduction to *mRMR*, read my article on **Towards Data Science**: [“MRMR” Explained Exactly How You Wished Someone Explained to You](https://towardsdatascience.com/mrmr-explained-exactly-how-you-wished-someone-explained-to-you-9cf4ed27458b).
Also, this article describes an example of *mRMR* used on the world famous **MNIST** dataset: [Feature Selection: How To Throw Away 95% of Your Data and Get 95% Accuracy](https://towardsdatascience.com/feature-selection-how-to-throw-away-95-of-your-data-and-get-95-accuracy-ad41ca016877).
*mRMR* was born in **2003**, this is the original paper: [Minimum Redundancy Feature Selection From Microarray Gene Expression Data](https://www.researchgate.net/publication/4033100_Minimum_Redundancy_Feature_Selection_From_Microarray_Gene_Expression_Data).
Since then, it has been used in many practical applications, due to its simplicity and effectiveness.
For instance, in **2019**, **Uber** engineers published a paper describing how they implemented MRMR in their marketing machine learning platform [Maximum Relevance and Minimum Redundancy Feature Selection Methods for a Marketing Machine Learning Platform](https://eng.uber.com/research/maximum-relevance-and-minimum-redundancy-feature-selection-methods-for-a-marketing-machine-learning-platform/).
Raw data
{
"_id": null,
"home_page": "https://github.com/smazzanti/mrmr",
"name": "mrmr-selection",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "",
"author": "Samuele Mazzanti",
"author_email": "mazzanti.sam@gmail.com",
"download_url": "",
"platform": null,
"description": "<p align=\"center\">\n<img src=\"./docs/img/mrmr_logo_white_bck.png\" alt=\"drawing\" width=\"450\"/>\n</p>\n\n## What is mRMR\n\n*mRMR*, which stands for \"minimum Redundancy - Maximum Relevance\", is a feature selection algorithm.\n\n## Why is it unique\n\nThe peculiarity of *mRMR* is that it is a **minimal-optimal** feature selection algorithm. <br/>\nThis means it is designed to find the smallest relevant subset of features for a given Machine Learning task.\n\nSelecting the minimum number of useful features is desirable for many reasons:\n- memory consumption,\n- time required,\n- performance,\n- explainability of results.\n\nThis is why a minimal-optimal method such as *mrmr* is often preferable.\n\nOn the contrary, the majority of other methods (for instance, Boruta or Positive-Feature-Importance) are classified as **all-relevant**, \nsince they identify all the features that have some kind of relationship with the target variable.\n\n## When to use mRMR\n\nDue to its efficiency, *mRMR* is ideal for practical ML applications, \nwhere it is necessary to perform feature selection frequently and automatically, \nin a relatively small amount of time.\n\nFor instance, in **2019**, **Uber** engineers published a paper describing how they implemented \n*mRMR* in their marketing machine learning platform [Maximum Relevance and Minimum Redundancy Feature Selection Methods for a Marketing Machine Learning Platform](https://eng.uber.com/research/maximum-relevance-and-minimum-redundancy-feature-selection-methods-for-a-marketing-machine-learning-platform/).\n\n## How to install this package\n\nYou can install this package in your environment via pip:\n\n<pre>\npip install mrmr_selection\n</pre>\n\nAnd then import it in Python through:\n\n<pre>\nimport mrmr\n</pre>\n\n## How to use this package\n\nThis package is designed to do *mRMR* selection through different tools, depending on your needs and constraints.\n\nCurrently, the following tools are supported (others will be added):\n- **Pandas**\n- **Polars**\n- **Spark**\n- **Google BigQuery**\n\nThe package has a module for each supported tool. Each module has *at least* these 2 functions:\n- `mrmr_classif`, for feature selection when the target variable is categorical (binary or multiclass).\n- `mrmr_regression`, for feature selection when the target variable is numeric.\n\nLet's see some examples.\n\n#### 1. Pandas example\nYou have a Pandas DataFrame (`X`) and a Series which is your target variable (`y`).\nYou want to select the best `K` features to make predictions on `y`.\n\n```python\n# create some pandas data\nimport pandas as pd\nfrom sklearn.datasets import make_classification\nX, y = make_classification(n_samples = 1000, n_features = 50, n_informative = 10, n_redundant = 40)\nX = pd.DataFrame(X)\ny = pd.Series(y)\n\n# select top 10 features using mRMR\nfrom mrmr import mrmr_classif\nselected_features = mrmr_classif(X=X, y=y, K=10)\n```\n\nNote: the output of mrmr_classif is a list containing K selected features. This is a **ranking**, therefore, if you want to make a further selection, take the first elements of this list.\n\n#### 2. Polars example\n\n```python\n# create some polars data\nimport polars\ndata = [(1.0, 1.0, 1.0, 7.0, 1.5, -2.3), \n (2.0, None, 2.0, 7.0, 8.5, 6.7), \n (2.0, None, 3.0, 7.0, -2.3, 4.4),\n (3.0, 4.0, 3.0, 7.0, 0.0, 0.0),\n (4.0, 5.0, 4.0, 7.0, 12.1, -5.2)]\ncolumns = [\"target\", \"some_null\", \"feature\", \"constant\", \"other_feature\", \"another_feature\"]\ndf_polars = polars.DataFrame(data=data, schema=columns)\n\n# select top 2 features using mRMR\nimport mrmr\nselected_features = mrmr.polars.mrmr_regression(df=df_polars, target_column=\"target\", K=2)\n```\n\n#### 3. Spark example\n\n```python\n# create some spark data\nimport pyspark\nsession = pyspark.sql.SparkSession(pyspark.context.SparkContext())\ndata = [(1.0, 1.0, 1.0, 7.0, 1.5, -2.3), \n (2.0, float('NaN'), 2.0, 7.0, 8.5, 6.7), \n (2.0, float('NaN'), 3.0, 7.0, -2.3, 4.4),\n (3.0, 4.0, 3.0, 7.0, 0.0, 0.0),\n (4.0, 5.0, 4.0, 7.0, 12.1, -5.2)]\ncolumns = [\"target\", \"some_null\", \"feature\", \"constant\", \"other_feature\", \"another_feature\"]\ndf_spark = session.createDataFrame(data=data, schema=columns)\n\n# select top 2 features using mRMR\nimport mrmr\nselected_features = mrmr.spark.mrmr_regression(df=df_spark, target_column=\"target\", K=2)\n```\n\n#### 4. Google BigQuery example\n\n```python\n# initialize BigQuery client\nfrom google.cloud.bigquery import Client\nbq_client = Client(credentials=your_credentials)\n\n# select top 20 features using mRMR\nimport mrmr\nselected_features = mrmr.bigquery.mrmr_regression(\n bq_client=bq_client,\n table_id='bigquery-public-data.covid19_open_data.covid19_open_data',\n target_column='new_deceased',\n K=20\n)\n```\n\n\n## Reference\n\nFor an easy-going introduction to *mRMR*, read my article on **Towards Data Science**: [\u201cMRMR\u201d Explained Exactly How You Wished Someone Explained to You](https://towardsdatascience.com/mrmr-explained-exactly-how-you-wished-someone-explained-to-you-9cf4ed27458b).\n\nAlso, this article describes an example of *mRMR* used on the world famous **MNIST** dataset: [Feature Selection: How To Throw Away 95% of Your Data and Get 95% Accuracy](https://towardsdatascience.com/feature-selection-how-to-throw-away-95-of-your-data-and-get-95-accuracy-ad41ca016877).\n\n*mRMR* was born in **2003**, this is the original paper: [Minimum Redundancy Feature Selection From Microarray Gene Expression Data](https://www.researchgate.net/publication/4033100_Minimum_Redundancy_Feature_Selection_From_Microarray_Gene_Expression_Data).\n\nSince then, it has been used in many practical applications, due to its simplicity and effectiveness.\nFor instance, in **2019**, **Uber** engineers published a paper describing how they implemented MRMR in their marketing machine learning platform [Maximum Relevance and Minimum Redundancy Feature Selection Methods for a Marketing Machine Learning Platform](https://eng.uber.com/research/maximum-relevance-and-minimum-redundancy-feature-selection-methods-for-a-marketing-machine-learning-platform/).\n\n\n",
"bugtrack_url": null,
"license": "GNU General Public License v3.0",
"summary": "minimum-Redundancy-Maximum-Relevance algorithm for feature selection",
"version": "0.2.8",
"project_urls": {
"Homepage": "https://github.com/smazzanti/mrmr"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "a9f4b09243bbcd133952f393b54924553c4a21ab2f4e4a7a083050f25afe5fcb",
"md5": "c11d849bfab6d600e7f78838cec4a3c4",
"sha256": "70723070d516802397e467075366dc3be07117de6378deee6821d5aba2686aa4"
},
"downloads": -1,
"filename": "mrmr_selection-0.2.8-py3-none-any.whl",
"has_sig": false,
"md5_digest": "c11d849bfab6d600e7f78838cec4a3c4",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 15149,
"upload_time": "2023-06-30T19:34:38",
"upload_time_iso_8601": "2023-06-30T19:34:38.977974Z",
"url": "https://files.pythonhosted.org/packages/a9/f4/b09243bbcd133952f393b54924553c4a21ab2f4e4a7a083050f25afe5fcb/mrmr_selection-0.2.8-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-06-30 19:34:38",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "smazzanti",
"github_project": "mrmr",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "mrmr-selection"
}
JSON
