Real-Time Big Data Analytics with R
summary
Shows how R plus Revolution R builds a real-time ad-click predictor: a logistic model trained on 1 million rows with 5% positive clicks (training run ~4 seconds) and deployed so scoring a handful of candidate ads returns probabilities in 0.01 seconds, letting the system pick the top ad. Uses Hadoop for the data lake, exports scoring rules via PMML or runs R on a DeployR server, supports continuous model refresh, and at scale generates 5 billion scores per day for retailers.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:52Revolution Analytics Microsoft
- 9:07R language history
- 16:29Why R for data science
- 23:52Predictive analytics ad targeting
- 27:22Real-time analytics defined
- 34:17Faster model development
- 42:21Live demo model in R
- 48:15Why you saw that specific ad on the webpage
- 51:48Learning R resources
- 52:29Q&A R in SQL Server
related talks
27mLarge Scale Response Prediction to Serve Ads on LinkedIn
37mBig Data Analytics, Machine Learning, Real Time Processing
36mML: Predictive analytics in Ad Engineering
32mMachine Learning using Open Source R
11mScaling R with Spark
38mRealtime Data Activation - Analytics, Intelligence & Decision Making
29mHow to automate the deployment of ...
41mFrom 5 PB of Daily Logs to Thousands of ML Models #bbuzz