Predicting Customer Behavior Using Logistic
Regression
Mansour Ahmady Phoulady
June 18, 2024
Mansour Ahmady Phoulady Predicting Customer Behavior Using Logistic Regression June 18, 2024 1/5
Scenario and Data
Scenario: Predict if a customer will make a purchase on the 11th day
based on their past behavior.
Data Overview:
Day Made Purchase? Amount Spent ($) Promotional Day?
1 1 25 1
2 0 0 0
3 1 30 0
4 1 45 1
5 0 0 0
6 1 20 0
7 0 0 0
8 1 50 1
9 0 0 0
10 1 40 1
Table: Purchasing Data for 10 Days
Mansour Ahmady Phoulady Predicting Customer Behavior Using Logistic Regression June 18, 2024 2/5
Model and Coefficients
Logistic Regression Model:
Logit(p) = β0 + β1 x1 + β2 x2
Where x1 is the amount spent yesterday, and x2 is whether it was a
promotional day.
Assumed Coefficients:
β0 = −1.0 (Intercept)
β1 = 0.05 (Effect of spending)
β2 = 1.0 (Effect of promotion)
Mansour Ahmady Phoulady Predicting Customer Behavior Using Logistic Regression June 18, 2024 3/5
Prediction for Day 11: Assuming
35spentandpromotionalday .Logit(p) = −1.0 + 0.05 × 35 + 1.0 × 1
Logit(p) = 1.75
Convert logit to probability:
p
= e 1.75 ≈ 5.75
1−p
5.75
p= ≈ 0.852
1 + 5.75
Probability of Purchase: Approximately 85.2%
Mansour Ahmady Phoulady Predicting Customer Behavior Using Logistic Regression June 18, 2024 4/5
Conclusion
This model predicts a high likelihood of the customer making a purchase
on the 11th day, especially if promotional activities are involved and recent
spending is high.
Mansour Ahmady Phoulady Predicting Customer Behavior Using Logistic Regression June 18, 2024 5/5
