Skip to main content

Real-time machine learning platform solutions

Through convenient visual operation, batch management and real-time synchronization of multi-source massive data can be efficiently completed. At the same time, the downstream can seamlessly connect with multiple business scenarios to support the application of business intelligence and artificial intelligence

Machine learning applications - personalized recommendation/risk control/marketing system

  • User Feature: Includes basic attributes of the user, historical behavior of the user and recent real-time behavior
  • Item Feature: Various attributes of an item, including discrete attributes and statistical value continuous attributes
  • User interaction feedback: that is, the label in the algorithm model

Build real-time training samples with LakeSoul

  • Multi-stream real-time update, three-stream in one
  • Use primary key to automatically extend the Schema, efficient MOR, no need to join
  • Downstream seamless docking with machine learning frameworks for real-time training of AI models

Join the community and share data intelligence