Content

Target customers: customers that have requirements for offline training and online real-time inference. Common scenarios include search, recommendation, and advertising, anomaly detection, and financial risk control. Features released: Machine Learning Designer allows you to deploy an offline pipeline that is used for data preprocessing, feature engineering, and model prediction to EAS. When you perform data prediction in an online environment, you can use the data preprocessing and feature engineering logic that is used in offline training. You do not need to process the data before you feed it to a model. This ensures the consistency of data processing.