Predicting Electricity Distribution Feeder Failures Using Boosting and Online Learning

author: Marta Arias, Universitat Politècnica de Catalunya

Categories

Top: Computer Science: Machine Learning: Boosting

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Predicting Electricity Distribution Feeder Failures using Machine Learning
2:24	Overview of the Talk
3:12	The Electrical System
5:04	Electricity Distribution: Feeders
5:52	Problem
7:59	Our Solution: Machine Learning
9:20	New York City
9:41	Some facts about feeders and failures pt 1
10:28	Some facts about feeders and failures pt 2
11:08	Some facts about feeders and failures pt 3
12:03	Feeder data
13:23	Feeder Ranking Application
13:54	Application Structure
14:08	Goal: rank feeders according to likelihood to failure
14:32	Overview of the Talk
15:11	(pseudo) ROC pt 1
16:12	(pseudo) ROC pt 2
16:57	(pseudo) ROC pt 3
17:23	Some observations about the (p)ROC
19:16	MartiRank pt 1
20:23	MartiRank pt 2
22:30	MartiRank pt 3
23:26	Using MartiRank for real-time ranking of feeders
24:56	Performance Metric
26:20	Performance Metric Example
26:43	How to measure performance over time
27:17	MartiRank Comparison: training every 2 weeks
29:09	Using MartiRank for real-time ranking of feeders
30:50	Overview of the Talk
32:05	Learning from expert advice pt 1
32:38	Learning from expert advice pt 2
33:41	Weighted Majority Algorithm [Littlestone & Warmuth ‘88]
35:00	In our case, can’t use WM directly
35:22	Dealing with ranking vs. binary classification
35:48	Dealing with a moving set of experts
37:36	Other parameters
38:45	Performance
39:28	Failures’ rank distribution
40:01	Daily average rank of failures
40:37	Other things that I have not talked about but took a significant amount of time
42:12	Current Status
42:58	Related work-in-progress
46:50	Other related projects within collaboration with Con Edison
47:50	Acknowledgments

Lecture rating

People found this lecture:
Worth seeing
because it is:
Valuable and informative
Well presented
Easily understandable
Acceptably recorded
You need to login to cast your vote.

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.