Live: Machine Learning Enablement on z/OS with DB2 and zSpark Integration

Project and Program: Information Management, Database

Tags: Proceedings, SHARE Providence 2017, 2017

A tight integration with Apache Spark on z/OS platform is essential. IBM Machine Learning for z/OS is an end-to-end enterprise machine learning platform that helps you create, train, and deploy machine learning models to extract value from your mission critical data on IBM z System while keeping the data where it resides. It utilizes IBM z/OS Platform for Apache Spark as the core machine learning runtime engine in the backend. This presentation will demonstrate the user interface of IBM Machine Learning from the "data scientists/data analyst", "model management" and "administrator dashboard" view of the product.
In the "data scientist/data analyst" view, data scientists can create, edit and share their applications with an integrated Jupyter Notebook, one of the most popular open source tools for model creation. Visual model builder, another model creation tool, is a guided model creation wizard that allows Data scientists to create Spark ML pipeline models without any programming skill. This method of model creation uses SPARK ingestion service for data ingestion and SPARK pipeline service to create, train, evaluate performance and save model. In "model management" view, application developers can use IBM Machine Learning for z/OS to manage models, create deployments (deploy models) and monitor model performance. In the "administrator dashboard" view, users can monitor their system resources, manage running kernels, and modify user roles and privileges.-Luan Bui-IBM; Tejaswini Karra-IBM

Back to Proceedings File Library