Skip to main content

This package provides helper utilities for machine learning tasks. One major utility is calculation of weight of evidence

Project description

Machine Learning Helper

This package usage multiple algorithms and parameters to accomodate different set of use cases to help in creating multiple machine learning algorithms.

1.0 woe (Weight of Evidence):

This function will help to calculate Weight of Evidence and Information Value, the charts can be displayed and coarse classing can also be done.

1.1 Parameters:


  • max_bin: int Maximum number of bins for numeric variables. The default is 10
  • iv_threshold: float Threshold value for Information Value. Variables with higher than threshold will be considered for transformation
  • ignore_threshold: Boolean This parameter controls whether the defined threshold should be considered or ignored. The default is 'True'

1.2 Returns:

DataFrame having weight of evidence of each column along with the target variable


1.3 Approach:

  1. Create an instance of woe my_woe = woe()

  2. Call fit method on the defined object by passing on dataframe and the target variable name my_woe.fit(df,target)

  3. Call the transform method transformed_df = my_woe.transform()


Example

Create Sample DataFrame

from mlh import woe
import pandas as pd
import numpy as np
import random

seed=1456
np.random.seed(seed)
random.seed(seed)
rows = 1000
y = random.choices([0,1],k=rows,weights=[.7,.3])
x1 = random.choices(np.arange(20,40),k=rows)
x2 = np.random.randint(1000,2000,size=rows)
x3 = random.choices(np.arange(1,100),k=rows)
x4 = random.choices(['m','f','u'],k=rows)
x5 = random.choices(['a','b','c','d','e','f','g','h'],k=rows)
df = pd.DataFrame({'y':y,'x1':x1,'x2':x2,'x3':x3,'x4':x4,'x5':x5})
df.head()

Fitting and prediction

Create Instance of Weight of Evidence Package

my_woe = woe()

Fit the data with created instance

my_woe.fit(df,'y')

Display the relevant charts

my_woe.getWoeCharts()

Merge values of X3 Variable at 1 and 2 indices using the Weight of Evidence chart from the first Iteration

my_woe.reset_woe(2,(1,2),1)

Get latest Iteration Information Value

my_woe.get_IV()

Replace the original values in the Dataframe with Weight of Evidence

transformed_df = my_woe.transform()

Project details


Download files

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

Filename, size & hash SHA256 hash help File type Python version Upload date
mlh-0.0.2-py3-none-any.whl (12.0 kB) Copy SHA256 hash SHA256 Wheel py3
mlh-0.0.2.tar.gz (10.6 kB) Copy SHA256 hash SHA256 Source None

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page