Making Predictions

Marketing Analytics: Predicting Customer Churn in Python

Mark Peterson

Director of Data Science, Infoblox

(Supervised) Machine Learning Primer

Goal: Predict whether or not a customer will churn
Target Variable: 'Churn'
Supervised Machine Learning
Learn from historical (training) data to make new predictions

scikit-learn logo

Model selection

Which model to use?
... it depends!
In this course: Experiment with several models
To learn about their inner workings: Check out other DataCamp courses

Model selection

Logistic regression: Good baseline
- Offers simplicity and interpretability
- Cannot capture more complex relationships
Random forests
Support vector machines

Training your model

from sklearn.svm import SVC

svc = SVC()

svc.fit(telco[features], telco['target'])

SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,
  decision_function_shape='ovr', degree=3, gamma='scale', kernel='rbf',
  max_iter=-1, probability=False, random_state=None, shrinking=True,
  tol=0.001, verbose=False)

Making a prediction

prediction = svc.predict(new_customer)

print(prediction)

[0]

Let's practice!

Marketing Analytics: Predicting Customer Churn in Python