Itaewon

My personal python library to speed-up my workflow.

Also, i found myself switching to R and Python quite frequently. This is my attempt to minimize that.

The implementation is quite simple, since i'm simply running R scripts in the background.

Sample usage is shown below.

Installation

# !pip install itaewon==0.1.4

import pandas as pd
import numpy as np
import warnings

warnings.filterwarnings('ignore')

data = pd.DataFrame()
error = np.random.normal(0,5,300)
data['X'] = np.random.normal(5,2.6,300)
data['Z'] = np.random.binomial(1,0.3,300)
data['Y'] = 10*data['X'] + 2*data['Z'] + error

Testing RModels

from itaewon.rmodels import RModels

formula = 'Y ~ X + Z'
rmodel = RModels(data,formula)
model = rmodel.linear_regression()
rmodel.infer_linear(model)

   Residuals       
 Min.   :-14.6393  
 1st Qu.: -3.4848  
 Median : -0.1583  
 Mean   :  0.0000  
 3rd Qu.:  3.3146  
 Max.   : 16.2795  

$coefficients
             Estimate Std. Error   t value      Pr(>|t|)
(Intercept) -1.109010  0.6718716 -1.650627  9.987172e-02
X           10.163182  0.1150433 88.342270 2.862585e-215
Z            1.426163  0.6222261  2.292033  2.260318e-02


$sigma
[1] 4.922223


$r.squared
[1] 0.9634044


$adj.r.squared
[1] 0.963158


fstatistic 3909.3683074439355
p_value [1] 4.668023e-214

new_data = pd.DataFrame()
new_data['X'] = [0,1,2,3]
new_data['Z'] = [1,0,0,1]

rmodel.predict(model,new_data,True,'confidence')

	Predicted	Lower	Upper
0	0.317153	-1.233912	1.868218
1	9.054173	7.921668	10.186677
2	19.217355	18.258526	20.176184
3	30.806700	29.671831	31.941568

rmodel.predict(model,new_data,True,'prediction')

	Predicted	Lower	Upper
0	0.317153	-9.493093	10.127399
1	9.054173	-0.698657	18.807003
2	19.217355	9.483164	28.951546
3	30.806700	21.053595	40.559804

rmodel.predict(model,new_data)

	Predicted
0	0.317153
1	9.054173
2	19.217355
3	30.806700

Testing the custom module

from itaewon.custom import Custom

from sklearn.linear_model import LinearRegression, LogisticRegression

lr = LinearRegression()

custom = Custom(data[['X','Z']],data['Y'],True,lr)
custom.custom()

RMSE for LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False) : 4.752017508515921

lr = LogisticRegression()

custom = Custom(data[['X','Y']],data['Z'],False,lr)
print(custom.custom()[0])
print(custom.custom()[1])

[[70  0]
 [29  0]]
              precision    recall  f1-score   support

           0       0.71      1.00      0.83        70
           1       0.00      0.00      0.00        29

    accuracy                           0.71        99
   macro avg       0.35      0.50      0.41        99
weighted avg       0.50      0.71      0.59        99