Josh Herbach: PLANET, MapReduce, and Decision Trees
The Association for Computing MachineryRate It
- Info
- Bio
- Full Program
- Highlights
- Transcript
- Download
- More
Chapters
-
Watch Full Program
01 hr 06 min 20 sec
- 01.
-
How Decision Trees Work
03 min 32 sec
- 02.
-
Building a Decision Tree
04 min 51 sec
- 03.
-
Problems with Decision Trees
02 min 57 sec
- 04.
-
Using Parallel Processing and MapReduce
04 min 32 sec
- 05.
-
Google's PLANET Decision Tree Trainer
03 min 42 sec
- 06.
-
The Role of PLANET's Master Process
03 min 15 sec
- 07.
-
Initializing MapReduce
02 min 57 sec
- 08.
-
Mapping to Find the Best Split
04 min 24 sec
- 09.
-
Reducing the Mapping Function
04 min 22 sec
- 10.
-
Other Aspects of the MapReduce Process
04 min 03 sec
- 11.
-
Bagging, Boosting, and Ensemble Modeling
04 min 17 sec
- 12.
-
Addressing Performance Issues
05 min 21 sec
- 13.
-
Sample Results
04 min 15 sec
- 14.
-
Conclusions
02 min 42 sec
- 15.
-
Q
00 min 04 sec
- 16.
-
Q1: Implementing PLANET Without MapReduce
01 min 33 sec
- 17.
-
Q2: Average Runtimes
01 min 14 sec
- 18.
-
Q3: Using PLANET to Predict Bounce Rate
01 min 41 sec
- 19.
-
Q4: Addressing Missing Data
01 min 37 sec
- 20.
-
Q5: Advantages of MapReduce Over Sampling
00 min 41 sec
- 21.
-
Q6: Handling Sparse Data
01 min 08 sec
- 22.
-
Q7: Selecting the Final Result
01 min 19 sec
- 23.
-
Q8: Other Ensemble Techniques
00 min 50 sec
- 24.
-
Q9: The Benefits of Optimization
00 min 53 sec
-
Watch Full Program
01 hr 06 min 20 sec
1 Comments
Subscribe to this Conversation 
Advertisement
Browse Programs by Tag
Arts
Books
Climate Change
Conflict
Credit Crunch
Defense
Democratization
Economic Policy
Education
Election 2008
Emerging Markets
Energy
Entertainment
Entrepreneurship
Environment
Film
Food
Globalization
Health & Wellness
Healthcare
History
Internet
Investing
Journalism
Law
Medicine
Music
Muslim World
Natural Sciences
New Media
Personal Technology
Philanthropy
Religion
Social Issues
Social Sciences
Sustainability
Terrorism
The Bailout
U.S. Foreign Policy
White House 2009
Advertisement
