StatsFLow Library

Functionality of the StatsFLow

• This Python package consist of various statistical modules and function that can be very useful for data preprocessing,data transformation,EDA and seamlessly improves the performance of Ml-Pipeline.

Usage

• Make sure you have Python installed in your system.
• Run Following command in the CMD.

  pip install statsflow

Example for probplot_transform() function

# importing the data  
df=pd.read_csv("data/UCI_Credit_Card.csv")
df.head()

#import probplot_transform function
from statsflow import probplot_transform

# apply the probplot_transform function

df['bill']=probplot_transform(feature=df['LIMIT_BAL'],trans_method='default')

Apply the probplot_transform function with various transformation['log','box_cox','recip','sqrt','expo']

#Exponential transformation
df['bill_expo']=probplot_transform(feature=df['LIMIT_BAL'],trans_method='expo')

#Square root transformation
df['bill_sqrt']=probplot_transform(feature=df['LIMIT_BAL'],trans_method='sqrt')

# Box-cox  transformation
df['bill_box_cox']=probplot_transform(feature=df['LIMIT_BAL'],trans_method='box_cox')

# Logarithmic Transformation
df['bill_log']=probplot_transform(feature=df['LIMIT_BAL'],trans_method='log')

#Reciprocal Transformation 
df['bill_log']=probplot_transform(feature=df['LIMIT_BAL'],trans_method='recip')

Note

-For Initial version i have only include the probplot_transform function.soon i will try to implement new statistical functions.

Probplot Transform Function:

Description:

This function used to interpret whether the given data follows normal distribution by using histogram and probplot. if the data not follows normal distribution then with the help of the trans_method param it can transform the data using Box-Cox,Logarithmic,Reciprocal,Square Root,Exponential Transformation.

Args:

[feature(pd.Series),trans_method('default','log','box_cox','recip','sqrt','expo')]