Introduction to Machine Learning

on YouTube ↗40m

" summary

Covers building a product recommender using ALS alternating least squares implemented with Spark ML + Scala, training on user-item “likes” and producing separate user and item factor matrices instead of a full prediction matrix. Uses PredictionIO, Heroku Postgres and Heroku Connect to run a demo called Dream House, shows unit tests that assert factor counts, RMSE drops with higher rank and more iterations, and prints actual recommendation scores — the demo trains and predicts successfully on the sample data.

AI-written summary · about

" chapters (23)

1. 0:00Introduction to ML
2. 1:21How children learn
3. 2:03Machine learning cycle
4. 3:08Machine learning as compression
5. 3:51Machine learning use cases
6. 5:57Collaborative filtering
7. 6:39Product recommendation demo
8. 7:43Tech stack overview
9. 8:04Product recommendation basics
10. 12:37Alternating least squares
11. 16:29Model training parameters
12. 17:18Model updates over time
13. 18:21Code walkthrough
14. 21:07Training method in Spark
15. 24:39Testing the model
16. 27:32Examining model features
17. 29:39Making predictions
18. 31:26Prediction method in Spark
19. 32:07Prediction test results
20. 34:36Model tuning
21. 37:24Overfitting and resources
22. 38:06Code and resources
23. 38:56Q&A

" related talks

• 31mCreating recommender systems with Spark, Scala and Prediction.IoFlorian Krause · code.talks 2017
• 37mBuild a Recommender System in Apache Spark and Integrate it Using AkkaWillem Meints · Scala Days Berlin 2016
• 48mBuild a Recommender System in Apache Spark and Integrate It Using AkkaWillem Meints · Scala Days Berlin 2016
• 40mBuilding your own Distributed Streaming Recommendation EngineRui Vieira · Berlin Buzzwords 2018
• 32mLeveraging recommender systems to personalise search resultsSoraya Hausl · PyData London 2017
• 58mRecommender Systems – A WalkthroughArtus Krohn-Grimberghe · PASS Business Analytics Conference 2013
• 29mThe Challenges of Building a Personalized Recommendation System in E-CommerceP. Poznic · Data Science Conference 4.0
• 33mPeople who liked this talk also liked ... Building Recommendation...Ryan Weald · LA RubyConf 2013