BDT

Gradient Boosting

NTrees and Shrinkage is the most important variables, basically, if not tuned, it will lead to overtraining.

But if we turn on the RandomizedTrees, for each tree, we only take a few of features, it will help a lot in improving overtraining condition.

Finding maximum/minimum

Evolution Strategy

https://blog.openai.com/evolution-strategies/#content

Smoothing

BoxCar

https://terpconnect.umd.edu/~toh/spectrum/Smoothing.html

http://iopscience.iop.org/article/10.1088/0741-3335/27/7/004/meta

-- RongkunWang - 2018-01-21


This topic: Sandbox > WebPreferences > RongkunSandbox > RongkunSeriesOfFacts
Topic revision: r3 - 2018-03-28 - RongkunWang
 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback