Dealing with Continuous Variable in Logistic Regression?

Question

I Am trying to create a Logistic model and I have two columns which are important features but it has continuous values ... I have 7044 rows so creating dummies might not be optimal which will burst whole lot of columns ... What should be My approach ... Any suggestions ...?

   PaymentMethod          MonthlyCharges    TotalCharges    Churn
  Electronic check          29.85              29.85          No
   Mailed check             56.95             1889.5          No
    Mailed check            53.85             108.15          Yes

A logistic regression requires a binary dependent variable but continuous predictors are no problem. — 00schneider, Jun 30 '20 at 10:24

kjetil b halvorsen · Answer 1 · 2020-06-30T16:06:46.143

3

Continuous predictors are fine with logistic regression. But, linear effects is a strong assumption with logistic regression, so it might be a good idea to spline the continuous predictors. See Using splines to address non-linearity in logistic regression and its links.

Your question contains little details and context, so if you want more advice, please update the post with details.

edited Jun 30 '20 at 16:06

answered Jun 30 '20 at 13:32

kjetil b halvorsen

63,378
26
142
467

https://www.kaggle.com/blastchar/telco-customer-churn – Harish reddy Jul 01 '20 at 09:41

Dealing with Continuous Variable in Logistic Regression?

1 Answers1