MindAccess Application Subscribers


The following Release Notes are valid only for region China 1.

Predictive Learning

What's new?

  • As an entirely new offering, Predictive Learning Essentials package is now available in Simplified Chinese in region China 1.

  • Predictive Learning Essentials package is an entry-level package that is Python-centric and light-weight. It supports a small development and execution environment based on small, general-purpose compute.

  • Predictive Learning Essentials package allows a data scientist to build prediction models through machine learning techniques, enabling companies to optimize product quality, which reduces potential field failures and performance issues.

  • Predictive Learning Essentials features:

    • Data Science Workbench: Managed instances of Jupyter notebook are provided for exploration, training and visualizing models with Python 2/ Python 3 Kernels. Users can flexibly install and update open-source packages to utilize in their model development.

    • Workspace Out-of-the-box functions: Enables users to quickly explore, transform and analyze big data without the need for any additional scripting.

    • Rapidly build and execute predictive models.

Currently not available in China 1:

  • Persist Python libraries across restart (on Predictive Learning Essentials)

  • UTS reporting includes sub-tenant information for Predictive Learning Storage

  • Faster quota usage reporting for Predictive Learning Storage

  • Pre-filled-in sources information in PrL UI if possible

  • Import PI Data

  • S/M/L package to support:

    • Clustered Compute by Spark EMR

    • Multi-Coding Language Support

    • Zeppelin Notebook

    • R kernel for Predictive Learning S, M and L in Jupyter