204.2.4 Goodness of fit for Logistic Regression

How good our model is?

204-2-4-goodness-of-fit-for-logistc-regression-in-python

Link to the previous post :https://statinfer.com/204-2-3-multiple-logistic-regression/

Goodness of Fit for a Logistic Regression

There are quite a lot of methods to find how good a model is. However, we will stick to the most important measures applicable for a logistic regression model.

Classification Matrix
AIC and BIC
ROC & AUC

Out of these three we will go through Classification matrix or confusion matrix in this post and understand rest in upcoming posts.

Classification Table & Accuracy

Predicted / Actual	0	1
0	True Positive (TP)	False Positive (FP)
1	False Negative (FN)	True Negative (TN)

Also known as confusion matrix
$A c c u r a c y = \frac{(T P + T N)}{(T P + F P + F N + T N)}$

Practice : Confusion Matrix & Accuracy

Create confusion matrix for Fiber bits model(Model built in previous posts of this series)

In [25]:

###for using confusion matrix###
from sklearn.metrics import confusion_matrix
cm1 = confusion_matrix(Fiber[['active_cust']],predict1)
print(cm1)

[[29210 12931]
 [10183 47676]]

Find the accuracy value for fiber bits model

In [26]:

total1=sum(sum(cm1))
accuracy1=(cm1[0,0]+cm1[1,1])/total1
accuracy1

Out[26]:

0.76885999999999999

The next post is about multicollinearity and individual impact of variables in logistic regression.
Link to the next post : https://statinfer.com/204-2-5-multicollinearity-and-individual-impact-of-variables-in-logistic-regression/

3rd March 2017