/tensorflow-recommendation-wals

An end-to-end solution for website article recommendations based on Google Analytics data. The model is based on matrix-factorization using the WALS algorithm, in TensorFlow, trained on Cloud ML Engine. Recommendations are served using App Engine Flex with Cloud Endpoints. Orchestration is performed using Airflow, running on Cloud Composer or GKE.

Primary LanguagePythonApache License 2.0Apache-2.0

Watchers