Happy to Announce Our New Discussion Forum | Join Now

Question-Machine Learning


Company XYZ needs a churn prevention model to predict which customers will NOT renew their yearly subscription to the company’s service.The company plans to provide these customers with a promotional offer.A binary classification model that uses Amazon Machine learning is reqd.On what basis shld the binary classification model be built?

A.User profiles (age,gender,income,occupation)

B Last user session

C Each user time series events in past 3 months

D Quaterly results


Option A and C looks most important features for given use case, however Option A could be more appropriate for predicting customer churn.
Its very difficult to choose best feature unless and until we do feature engineering with sample dataset
AWS Blog on similar problem : https://aws.amazon.com/blogs/machine-learning/predicting-customer-churn-with-amazon-machine-learning/