Predicting the future as a service with Azure ML and R
summary
Walks through building a binary income classifier in Azure ML Studio using a 32k-row, 15-column dataset, a boosted decision tree, and an 80/20 split for train/test; shows scoring, evaluation metrics beyond accuracy, and hyperparameter tuning. Demonstrates embedding custom R (or Python) modules, turning the trained model into a deployed REST API with an auth key, and options to retrain or import gallery experiments; the demo runs end-to-end and the service returns scores for new records.
AI-written summary · about
chapters (28)
- 0:00Introduction
- 1:07Agenda
- 2:02What is machine learning
- 5:49How machine learning works
- 13:56How human brain works
- 18:14Unsupervised learning
- 19:11Unsupervised learning workflow
- 23:51You need to know your stuff
- 25:13What is Azure ML
- 26:02Demo
- 26:45New Experiment
- 29:39Split Data
- 31:39Train Model
- 33:45Score Model
- 34:45Predicted Labels
- 35:38Evaluate Model
- 36:21Run Model
- 37:18Decision Tree
- 40:06Accuracy
- 41:30Machine Learning
- 43:18Scripts
- 44:13Why use Azure ML
- 45:58What does it do
- 48:48Web Services
- 51:35Cortana
- 52:05Twitter Sentiment
- 53:47Text Data
- 58:15Summary
related talks
01h 05mAzure Machine Learning for the Developer
48mMachine Learning in Azure Databricks
53mImplementing AI using Azure Machine Learning Services
01h 01mCitizen data science
38mReal Life Customer Stories with AI
41mLeveraging Automated Machine Learning to enable anyone to develop Machine Learning solutions
37mModel Creation in Azure AutoML and Ingestion Through Excel
24mAzure ML, Azure Kubernetes Service, Model deployment