What features are generated after the Engine processes the data in a transactional customer churn prediction app?

This article explains the features that are generated after the Engine processes the data in a transactional customer churn prediction app.

Features are known attributes used as input by machine learning models to predict the unknown target.

For the transactional option in the customer churn template app, the AI & Analytics Engine automatically generates a number of useful features from the customer transactional dataset and customer information dataset. These features represent various customer behavior statistics over different periods of time.

Different types of aggregated features are generated over the selected time windows:

• Amount-based features: minimum, maximum, standard deviation, and total amount spent and refunded.

• Count-based features: number of transactions.

• Time-interval-based features: minimum, maximum, standard deviation, and average number of days between transactions.

• Recency features: days since last transaction.

The user can select the time windows used to generate these aggregated features when they are defining contributing factors.

Apart from aggregated features, additional demographic features from the customer info data are also included, if available.

Specifying two time windows. One is the most recent 30 days. The other is a 30-day range, 30 days ago.

As an example, consider the following credit-card transactions dataset and customer information dataset as input to the template:

Transaction dataset

Customer information dataset

Then, the following features will be generated by the engine. For the transaction activity based features, the suffixes such as _last_30d and _last_15d correspond to the time windows we have chosen while selecting the contributing factors: