Implement Machine Learning Using Oracle Data Miner LVC

Course Overview

• Learning Objectives
• Course Objectives
• Course Goals
• Target Audience
• Prerequisites
• Proposed Schedule
• Prerequisite Courses
• Summary

Fundamentals of Oracle Machine Learning

• Learning Objectives
• Data Science Landscape
• Terminology Consensus?
• What is Machine Learning?
• Real World Machine Learning Examples
• Examples of Machine Learning Use Cases
• Basic ML Terminology
• ML Approaches
• What is Machine Learning?
• ML Techniques Use Cases
• CRISP-DM – How to manage an ML Project
• CRISP-DM – The Iterative 6 Phases
• Traditional vs. Oracle Machine Learning/Predictive Analytics
• Move the Algorithms, Not the Data!
• Traditional ML vs. Oracle ML Cost Savings
• Oracle Machine Learning
• Oracle Machine Learning Interfaces to Oracle Database
• Oracle Machine Learning for SQL (OML4SQL)
• OML for SQL Model Build & SQL Apply
• Oracle Data Miner
• OML4R
• Oracle Machine Learning
• Oracle Machine Learning for Python
• Coming Soon! | AutoML – new with OML4Py
• The Evolution of the DBA/Database Developer Role
• What Autonomous Database means for DBAs
• Database Developer to Citizen Data Scientist Journey
• The Changing Role of the DBA: Motivation to Learn to Become or Support a Data
• Scientist!
• Summary

Introduction to Oracle Machine Learning UIs

• Learning Objectives
• Oracle Machine Learning Interfaces to Oracle Database
• Highlights of the Oracle Machine Learning API
• Oracle Machine Learning for SQL (OML4SQL)
• Oracle Machine Learning for SQL Features
• Advantages of OML4SQL in the Database Kernel
• Creating an OML4SQL Model
• Creating an OML Model Using PL/SQL API
• Executing an OML Model using SQL Functions
• Interfaces to OML4SQL: OML UIs
• Introduction to Oracle Data Miner
• Oracle Data Miner (ODMr) Architecture
• Connections for Machine Learning
• Setting up Oracle Data Miner
• Step 1: Create a Connection for ODMr User
• Step 2: Make the Data Miner UI Visible
• Step 3: Install the ODMr Repository
• ODMr Repository Installation Process
• Introducing the ODMr Interface
• Creating Oracle Data Miner Projects
• Creating Oracle Data Miner Workflows
• Building a Workflow
• Examining ODMr Nodes – Components
• Examining ODMr Nodes – Data Understanding
• Building a Sample Workflow – Add a Data Source
• Building a Sample Workflow – Add a Graph
• Building a Sample Workflow – Explore Data
• Examining ODMr Nodes – Data Preparation
• Building a Sample Workflow – Filter Data
• Examining ODMr Nodes – Modeling
• Building a Sample Workflow – Add a Classification Node
• Examining ODMr Nodes – Evaluation
• Building a Sample Workflow – Apply the Model
• Examining ODMr Nodes – Additional Nodes
• Completed Sample ODMr Workflow
• Running the ODMr Workflow
• Managing Workflows
• Deploying a Workflow – Deploy SQL Scripts
• Deploying a Workflow – Save SQL
• OML Deployment Use Cases – Query in Code
• OML Deployment Use Cases – Query in APEX
• OML Deployment Use Cases – Query in REST
• Types of Model Scoring In Queries
• Oracle Machine Learning Interfaces to Oracle Database
• OML Notebooks Architecture
• OML Notebooks Web-based UI
• Working with Oracle Machine Learning Notebooks
• OML Notebooks – Work Collaboratively
• OML Notebooks – Easily Access Data
• OML Notebooks – Deploy Models Inside the Database
• Oracle Data Miner or OML Notebooks?
• OML: Algorithm Cheat Sheet
• Summary
• Practice 3: Introduction to Oracle ML Tools

Using Classification Models

• Lesson Objectives
• ML Approaches
• Types of Supervised Learning
• Supervised Modeling
• Training vs Test Data
• How do you classify things?
• Classification: Use Cases
• Classification and Attribute Dependency
• Importance of the Target Attribute
• Classification: Scoring and Deployment
• Building a Simple Classification Model using ODMr
• Phase 1: Business Understanding – Background
• Phase 1: Business Understanding – Requirements
• Phase 1: Business Understanding – Museum Touch Screen Application
• Phase 1: Business Understanding – Summary of Objectives and Success
• Criteria
• Phase 1: Business Understanding – Sharp Questions
• Phase 1: Business Understanding – Initial Assessment of ML Techniques
• Phase 2: Data Understanding – Defined
• Phase 2: Data Understanding – Explore Data Node
• Titanic Data Set
• Phase 2: Data Understanding with Explore Data Node
• Phase 2: Data Understanding Target
• Phase 2: Data Understanding – Graph Node
• Phase 2: Data Understanding with Graph Data Node
• Phase 3: Data Preparation – Transforms Nodes
• Phase 3: Data Preparation – The Transform Node
• Phase 3: Data Preparation – Cleaning Null Values
• Phase 3: Data Preparation – Binning
• Phase 3: Data Preparation – Transformed Fields
• Phase 3: Data Preparation – SQL Query Node
• Phase 3: Data Preparation – Explore Data on Transformed Fields
• Phase 3: Data Preparation – Filter Columns Node
• Phase 3: Data Preparation – Attribute Importance Technique
• Phase 3: Data Preparation with Filter Columns Node
• Phase 3: Data Preparation Phase Complete
• Phase 3: Data Preparation Phase Complete – Dataset
• Phase 4: Modeling – Models Nodes
• Revisiting Classification: Modeling
• Phase 4: Modeling – Classification Build Node
• Phase 4: Modeling – Advanced Settings (Optional)
• Phase 4: Modeling – Training vs Test Dataset Settings
• Phase 4: Modeling – Building the Classification Model
• Phase 4: Modeling – Compare Model Performance
• Phase 4: Modeling – Comparing Test Results – Performance Tab
• Phase 4: Modeling – Comparing Test Results – Performance Matrix

Phase 4: Modeling – Comparing Test Results – Receiver Operating Characteristics

• (ROC)
• Some Algorithms are Transparent
• Phase 4: Modeling – Selecting the Decision Tree Algorithm
• Phase 4: Modeling Phase Complete
• Phase 5: Evaluation 1 of 2
• Phase 5: Evaluation 2 of 2
• Phase 6: Deployment – Model Operations Nodes
• Revisiting Classification: Scoring and Deployment
• Phase 6: Deployment – Scoring Dataset
• Phase 6: Deployment – Apply Node
• Phase 6: Deployment – Edit Apply Node
• Phase 6: Deployment – View Scoring Results
• Phase 6: Deployment – Deploy Apply Node for Real Time Scoring
• Phase 6: Deployment – View Apply Node SQL in SQL Worksheet
• Phase 6: Deployment – Demo Scoring Web UI (with Apply Node SQL)
• Traditional ML Process vs Automated OML Process
• Automated OML Process – Automatic Data Preparation (ADP)
• Titanic Use Case using Automated OML Process
• Automated OML Process – Model Performance
• Automated OML Process – Performance Matrix aka Confusion Matrix
• Traditional ML Process vs Automated OML Process – Performance
• Summary
• Practice 4: Overview

Using Regression Models

• Lesson Objectives
• Revisiting Supervised Learning – OML Algorithms
• Regression: Data 1 of 3
• Regression: Data 2 of 3 – Sample Data
• Regression: Data 3 of 3 – Training vs Test Data
• Regression: Modeling
• Regression: Scoring and Deployment
• Building a Regression Model using Oracle Data Miner
• Boston Housing Sample Dataset for Regression
• Phase 1: Business Understanding – Requirements
• Phase 2: Data Understanding – Data Nodes
• Phase 2: Data Understanding –Explore Data Node
• Phase 2: Data Understanding with Explore Data Node
• Phase 2: Data Understanding – ODMr Graph Node
• Phase 2: Data Understanding with Graph Data Node – Box Plot and Outliers
• Phase 2: Data Understanding with Graph Data Node – Scatter Plots and Correlations
• Phase 3: Data Preparation –Transform Nodes
• Phase 3: Data Preparation – Transform Node
• Phase 3: Data Preparation – Outlier Treatment
• Phase 3: Data Preparation – Transformed Fields
• Phase 3: Data Preparation – Filter Columns Node
• Phase 3: Data Preparation – Attribute Importance Technique
• Phase 3: Data Preparation with Filter Columns Node
• Phase 3: Data Preparation Phase Complete – Prepared Dataset
• Phase 3: Data Preparation Phase Complete
• Phase 4: Modeling – Models Nodes
• Revisiting Regression: Modeling
• Phase 4: Modeling – Regression Node
• Phase 4: Modeling – Regression Build Node
• Phase 4: Modeling – Advanced Settings (Optional)
• Phase 4: Modeling – Training vs Test Dataset Settings
• Phase 4: Modeling – Building the Regression Model
• Phase 4: Modeling – Compare Model Performance
• Phase 4: Modeling – Comparing Test Results – Performance Tab
• Phase 4: Modeling – Comparing Test Results – Residuals Tab
• Phase 4: Modeling – Selecting the SVM Algorithm
• Phase 4: Modeling Phase Complete
• Phase 5: Evaluation
• Phase 5: Evaluation Phase
• Phase 6: Deployment – Model Operations Nodes
• Phase 6: Deployment – Scoring Data set
• Phase 6: Deployment – ODMr Scoring Dataset
• Phase 6: Deployment – Apply Node
• Phase 6: Deployment – View Scoring Results
• Phase 6: Deployment – Deploy Apply Node for Real Time Scoring
• Phase 6: Deployment – View Apply Node SQL in SQL Worksheet
• Phase 6: Deployment Phase Complete – Deployment
• Revisiting Traditional ML process vs Automated OML Process
• Automated OML Process – Automatic Data Preparation (ADP)
• Summary
• Practice 5: Overview

Using Clustering Models

• Objectives
• Revisiting ML Techniques 1 of 2
• Revisiting Unsupervised Learning: OML Algorithms
• Clustering: Data 1 of 3
• Clustering: Data 2 of 3: Sample Data
• Clustering: Modeling
• Clustering: Scoring and Deployment
• Building a Clustering Model Using Oracle Data Miner
• Insurance Customers Dataset for Clustering
• Phase 1: Business Understanding – Requirements
• Phase 2: Data Understanding – ODMr Data Nodes
• Phase 2: Data Understanding – ODMr Explore Data Node
• Phase 2: Data Understanding with Explore Data Node
• Phase 3: Data Preparation – ODMr Transforms Nodes
• Phase 4: Modeling – ODMr Model Nodes
• Phase 4: Modeling – ODMr Clustering Node
• Phase 4: Modeling – Edit Clustering Build Node
• Phase 4: Modeling – k-Means Advanced Settings
• Phase 4: Modeling – O-Cluster Advanced Settings
• Phase 4: Modeling – Expectation Maximization Advanced Settings
• Phase 4: Modeling – Building the Clustering Model
• Phase 4: Modeling – ODMr Models Nodes
• Phase 4: Modeling – k-Means Model Results
• Phase 4: Modeling – k-Means Model Results – Tree Navigation
• Phase 4: Modeling – k-Means Model Results – Rules Tab
• Phase 4: Modeling – k-Means Model – Cluster Details Tab
• Phase 4: Modeling – k-Means Model – Cluster Compare Tab
• Phase 4: Modeling – k-Means Model – Compare Tab Rename Clusters
• Phase 4: Modeling – Select a Clustering Model
• Phase 5: Evaluation
• Phase 6: Deployment – ODMr Model Operations Nodes
• Phase 6: Deployment – ODMr Scoring Dataset
• Phase 6: Deployment – ODMr Apply Node
• Phase 6: Deployment – View Scoring Results
• Phase 6: Deployment Phase Complete – ODMr Deployment
• Revisiting Traditional ML Process vs Automated OML Process
• Automated OML Process: Automatic Data Preparation (ADP)
• Summary
• Practice 6: Overview

Using Anomaly Detection Models

• Objectives
• Revisiting Unsupervised Learning Algorithms
• Anomaly Detection: Data 1 of 2
• Anomaly Detection: Data 2 of 2 – Sample Data
• Anomaly Detection: Modeling
• Anomaly Detection: Scoring and Deployment
• Building an Anomaly Detection Model Using CRISP –DM and ODMr
• Phase 1: Business Understanding – Requirements
• Phase 1: Business Understanding – Initial Assessment of ML Techniques
• Phase 2: Data Understanding – Sample Dataset 1 of 2
• Phase 2: Data Understanding – Sample Dataset 2 of 2
• Phase 2: Data Understanding – ODMr Data Nodes
• Phase 2: Data Understanding – ODMr Explore Data Node
• Phase 2: Data Understanding with Explore Data Node
• Phase 3: Data Preparation – ODMr Transforms Nodes
• Phase 3: OML Automatic Data Preparation (ADP)
• Phase 4: Modeling – ODMr Model Nodes
• Revisiting Anomaly Detection: Modeling
• Phase 4: Modeling – ODMr Anomaly Detection Node
• Phase 4: Modeling – Edit Anomaly Detection Build Node
• Phase 4: Modeling – Building the Anomaly Detection Model
• Phase 5: Evaluation – ODMr Model Nodes
• Phase 5: Evaluation – Anomaly Detection Model Results
• Phase 5: Evaluation Phase Complete
• Phase 6: Deployment – ODMr Model Operation Nodes
• Phase 6: Deployment – ODMr Apply Node
• Phase 6: Deployment – View Scoring Results
• Phase 6: Deployment Phase Complete – ODMr Deployment
• Summary
• Practice 7: Overview