AppDev with Machine Learning and AutoML

on YouTube ↗31m

" summary

Builds a churn prediction for a fictional MovieStream service inside an Autonomous Database using the AutoML UI, which auto-selects algorithms, samples, features and tunes models. AutoML picks a Random Forest as best and the model reports 64% correct nonchurn, 28% true positives, 6% false positives and 1% missed churners. The model deploys to Oracle Machine Learning Services as a REST endpoint, returns a prediction in ~47 ms and scores 100-customer microbatches subsecond.

AI-written summary · about

" chapters (12)

1. 0:00Introduction to OML
2. 0:57Industry use cases
3. 2:50Database advantages
4. 4:55Movie streaming scenario
5. 6:47Automated machine learning
6. 10:47Setting up the experiment
7. 16:10Analyzing model results
8. 19:21Deployment options
9. 22:51APIs and flexibility
10. 27:09Embedded execution
11. 28:40Data and model monitoring
12. 30:09Roadmap and takeaways

" related talks

• 24mMachine Learning for Analytics TeamAshwath Raj, Rajiv Kumar Maskara · MachineCon 2022
• 22mPredict the future (or don't) with machine learning in applicationsMark Hornick · Oracle DatabaseWorld 2022
• 23mCustom-made games with machine learning and big dataÁfrica Periañez · Pocket Gamer Connects Seattle 2019
• 20mBuilding Intelligent Retention Engines: Real-Time Churn DetectionGaurav Sunkar · Conf42 Kube Native 2025
• 32mExplainable AI Explained for DevelopersWillem Meints, Joop Snijder · GOTO Amsterdam 2022
• 08mBroadening AI Adoption with AutoMLBernhard Suhm · ML Conf EU 2020
• 38mFantastic ML apps and how to build themMatthew Tovbin · Scale By the Bay 2017
• 31mIn-database machine learning: Automate, scale, deployMark Hornick · Oracle DatabaseWorld 2021