@Nizam you can use the adaptive gradient boosting algorithm for e-commerce recommendations.

It is a suitable use case because the algorithm provides higher predictive power, which can lead to more accurate and relevant recommendations for customers. However, keep in mind that adaptive gradient boosting models are more complex and have a lower transparency score compared to Bayesian models.

If transparency is a priority, you might want to consider using the default Bayesian modeling approach in Pega Customer Decision Hub.

As a Support Center moderator, I reviewed the answer and references for accuracy  

Adaptive gradient boosting overview > Advanced model settings

Adaptive analytics > Transcript