What do beta coefficients of Logistic Regression say ?

N number of statistical tools are available in market for making our life easy. One click and a super complex model on enormous data is prepared within nanoseconds.

However, no software is yet available  to translate these high end mathematical models into business insights… except one ...

Yes, Human mind !

Would you like to program your mind so that it can translate Logistic Regression model full of betas into layman and business language ? 

In order to program your mind, let's, first, learn few basic comparison methods using below mentioned literacy rate example :

Above examples have just made us "literate" about various comparison methods and few terms : odds and odds ratio.

Keep above comparison method in your mind. This is going to help you to interpret logistic regression result.

Interpretation of Logistic Regression Result 

Taking an example of telecom industry.We have shown sample (fictitious) data below: 

Below we have mentioned the logistic equation without getting into the details of the logistic model.

We have already covered Logistic Regression into detail in one of our previous articles:
Learn Logistic Regression

Continued ...

We have calculated the betas/coefficients and derived the equation.

You have calculated the probabilities of being churn of two scenarios.

Now think of comparison methods which we have learnt.

We would follow the same comparison methods here.

I know this is little complex, but try reading it two or three times with a break and if you still find yourself confused, do let me know. I will write it again to make it simper!


Enjoy reading our other articles and stay tuned with ...

Kindly do provide your feedback in the 'Comments' Section and share as much as possible.