Introduction to Predictive Learning¶
MindSphere Predictive (PrL) Learning enables data scientists to write or bring their own machine learning (ML) algorithms using statistical functions, transformations, and filtering to bring the most comprehensive and flexible means for accessing and working with both historical and near real-time data. PrL supports multiple data sources, such as:
- Integrated Data Lake (IDL) : Folders in IDL can be referenced from PrL.
- Internet of Things (IoT) : Asset data can be made available to PrL.
- Predictive Learning Storage : Larger Files like parquet files are supported using this internal storage of PrL.
- Data Exchange : Folder structures with smaller files with size less than 100MB can be uploaded to this internal storage.
PrL features allow users to:
- Develop, train, and execute models within environments
- Create Jupyter notebooks and save them locally while environments continue to run
- Use Python scripts to read data from APIs
- Bring locally build models for execution as Dockers
- Schedule the model execution
- Use variety of data sources for Input and Output
Roles and Permissions¶
Predictive Learning administrators create and edit roles and permissions via the Settings app on the MindSphere Launchpad. If the Settings app does not appear on the Launchpad, please contact your MindSphere administrator.
Any questions left?
Except where otherwise noted, content on this site is licensed under the MindSphere Development License Agreement.