ML practices in the industry
summary
Frames Big Data as the infrastructure layer — Hadoop/HDFS, NameNode, Sqoop/Flume, Pig, Hive, Zookeeper and Mahout — and data science as applied analytics using SAS/R/Python plus basic statistics and machine learning. Explains map reduce as split-and-merge work and argues Spark in-memory runs many jobs “100x” faster. Gives a concrete result: sending 100,000 mailers yields 1,000 buyers, targeting 20,000 by modeling yields 900 buyers and cuts CPA from ~$200 to ~$45. Says Big Data skills take ~2 months if you know SQL; analytics requires ~6 months of practice.
AI-written summary · about
chapters (10)
- 0:00Start
- 2:41Three Components of Data Science
- 5:42What Is Artificial Intelligence
- 9:35What Is Mapreduce
- 15:44Data Science
- 21:00What Are the Machine Learning Practices in the Industry
- 21:15Type of Machine Learning
- 26:18Classification Tree
- 29:49Regression Tree
- 30:20Logistic Regression
related talks
37mBig Data Analytics, Machine Learning, Real Time Processing
28mAI/ML for Non Data Scientists
01h 35mDeep Dive into the Big Data Landscape
44mData Science at Scale with Apache Spark
45mSolving Real & Big (Data) Problems Using Hadoop
35mPySpark : Combining Machine Learning & Big Data
24mImprove your marketing reach using ...
17mMachine Learning at the Speed of Light with Spark