Model Creation in Azure AutoML and Ingestion Through Excel
summary
Creates a classification model in Azure Machine Learning Studio using AutoML (runs many preprocessing + algorithm trials automatically), finds a stack ensemble as best with about 95% AUC weighted, and demonstrates deploying it as a web endpoint (container instance) and calling it from Excel via VBA. Demo uses a D&D character CSV, shows feature importances, ROC and confusion matrix, notes dwarf class mispredicted due to extreme class imbalance and a speed=25 feature quirk, and notes AutoML’s endpoint does not return per-class confidence.
AI-written summary · about
chapters (12)
- 0:00Introduction and agenda
- 0:33Overview of topics
- 1:49Machine learning basics
- 3:39What is AutoML?
- 4:43Azure ML studio overview
- 8:06Who should use AutoML?
- 9:52Problems AutoML solves
- 11:45Data upload and job setup
- 18:29Analyzing model results
- 24:27Machine learning workflow
- 27:38Deploying models to Excel
- 31:20Questions and conclusion
related talks
01h 05mAzure Machine Learning for the Developer
58mDeep Dive into Azure Automated Machine Learning
59mPredicting the future as a service with Azure ML and R
41mLeveraging Automated Machine Learning to enable anyone to develop Machine Learning solutions
53mImplementing AI using Azure Machine Learning Services
01h 01mAI and ML imbedded in Azure Data Platform
40mCloud-based Machine Learning Offerings - Beginner's Perspective
01h 14mAI of Things