Predicting Electricity Distribution Feeder Failures Using Boosting and Online Learning

Posted in Science on September 06, 2008


Predicting Electricity Distribution Feeder Failures Using Boosting and Online Learning

Lecture slides:

  • Predicting Electricity Distribution Feeder Failures using Machine Learning
  • Overview of the Talk
  • The Electrical System
  • Electricity Distribution: Feeders
  • Problem
  • Our Solution: Machine Learning
  • New York City
  • Some facts about feeders and failures
  • Feeder data
  • Feeder Ranking Application
  • Application Structure
  • Goal: rank feeders according to likelihood to failure
  • Overview of the Talk
  • (pseudo) ROC
  • Some observations about the (p)ROC
  • MartiRank
  • Using MartiRank for real-time ranking of feeders
  • Performance Metric
  • Performance Metric Example
  • How to measure performance over time
  • MartiRank Comparison: training every 2 weeks
  • Using MartiRank for real-time ranking of feeders
  • Overview of the Talk
  • Learning from expert advice
  • Weighted Majority Algorithm [Littlestone & Warmuth ‘88]
  • In our case, can’t use WM directly
  • Dealing with ranking vs. binary classification
  • Dealing with a moving set of experts
  • Other parameters
  • Performance
  • Failures’ rank distribution
  • Daily average rank of failures
  • Other things that I have not talked about but took a significant amount of time
  • Current Status
  • Related work-in-progress
  • Other related projects within collaboration with Con Edison
  • Acknowledgments

Author: Marta Arias, Universitat Politècnica De Catalunya

Watch Video

Tags: Science, Lectures, Computer Science, Machine Learning, VideoLectures.Net, Boosting