Oracle Analytics Cloud and Server Idea Lab

Welcome to the Oracle Analytics Community: Please complete your User Profile and upload your Profile Picture

Tight integration with ADW's Oracle Machine Learning

Delivered
42
Views
3
Comments

Organization Name

Oracle

Description

If OAC is connected to ADW and the data is in ADW, then the Explain, Outlier, etc, ML features should leverage Oracle Machne Learning's ML algorithms and perform all ML processing in-Database.  There are other tighter integrations such as auto-displaying OML's prediction probabilities as deciled probability buckets for bar charting and auto-displaying OML's Prediction_Details that could also be implemented/integrated.

 

Use Case and Business Need

More auto-insights capabilities and auto interactive viz and analysis of OML insights and predictions when the data is stored in an Oracle Database.  Deeper insights, predictions and more information harvested from data.

More details

The OAC + OML integration announced is a great start down this journey.   Want to see more!  Thanks!

Original Idea Number: 2af3fbebb2

Snap41.jpg

7
7 votes

Delivered · Last Updated

Comments

  • Can't comment here Charlie, but I hear you and raise you with microservice OML :) 

  • TimVlamis
    TimVlamis Rank 5 - Community Champion

    I agree whole-heartily with Charlie's comment. There are many of tables that are automatically created in Oracle database when certain functions from Oracle Machine Learning are run (Oracle Data Mining/OML4SQL). it would be great to see more auto-generated connections and visualizations in OAC. For example, when you run a hierarchical K-Means clustering algorithm, have the capability to produce a visualization of the tree and a linked visualization showing the importance of variables for a given selected node in the tree. There are many other examples from Data Miner of what would be useful on an auto-generation basis.

    It's the presentation and interpretation of the data from Oracle Machine Learning that will provide value to businesses. We can always produce tables of results that a data scientist can read, but ultimately, helping people on the business side understand and use the results is what will drive value. Of course, some training in how to understand complex machine learning models will be useful.