The 18th European Conference on Machine Learning (ECML) and the 11th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD)

Putting Things in Order: On the Fundamental Role of Ranking in Classification and Probability Estimation

author: Peter A. Flach, University of Bristol

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Putting things in order
0:17	Motivation
3:01	Motivation (2)
4:26	Motivation (3)
9:09	Outline
9:44	Health warning
11:13	Building models
12:06	Decision tree classifier (1)
13:22	Decision tree classifier (2)
13:47	Decision tree ranker
15:04	Decision tree probability estimator
15:57	Visualising ranking performance
18:00	Visualising ranking performance (2)
18:20	Visualising ranking performance
18:25	Visualising ranking performance (2)
19:16	Visualising ranking performance
20:04	Visualising ranking performance (2)
20:10	All possible tree labellings
22:22	Building the tree recursively (1)
23:19	Building the tree recursively (2)
23:23	Building the tree recursively (3)
23:57	Reordering ROC segments (1)
24:07	Reordering ROC segments (2)
25:01	Naive Bayes probability estimator
27:52	Naive Bayes ROC curve
29:17	Lexicographic ranking
29:46	Lexicographic ranking (2)
31:29	Odds ratio splitting criterion (1)
32:12	Odds ratio splitting criterion (2)
32:30	Odds ratio splitting criterion (3)
32:33	Odds ratio splitting criterion (4)
32:45	Cue: ProgRoc
37:33	Building models — summary
38:06	II Classification and ranking
39:17	Some notation
39:52	From a ranking to a ROC curve (1)
40:24	From a ranking to a ROC curve (2)
40:32	Machine Learning 101 Exam, Q42. AUC is ... (1)
41:01	Machine Learning 101 Exam, Q42. AUC is ... (2)
41:15	Machine Learning 101 Exam, Q42. AUC is ... (3)
41:35	Machine Learning 101 Exam, Q42. AUC is ... (4)
41:46	Machine Learning 101 Exam, Q42. AUC is ... (5)
41:57	Machine Learning 101 Exam, Q42. AUC is ... (6)
42:19	Machine Learning 101 Exam, Q42. AUC is ... (7)
43:06	Machine Learning 101 Exam, Q42. AUC is ... (8)
43:16	Machine Learning 101 Exam, Q42. AUC is ... (9)
43:40	From AUC to accuracy
45:21	From accuracy to AUC (1)
46:18	From accuracy to AUC (2)
46:19	From accuracy to AUC (3)
46:20	From accuracy to AUC (4)
46:21	From accuracy to AUC (5)
46:22	From accuracy to AUC (6)
46:26	From accuracy to AUC (7)
46:48	From accuracy to AUC (8)
46:57	From accuracy to AUC (9)
47:15	Classification and ranking — summary
47:38	III Ranking and probability estimation
48:17	Calibration
51:41	Calibration through the ROC convex hull
53:12	Alternative: logistic calibration
55:46	Decomposing the Brier score
57:39	Decomposing the Brier score (Flach & Matsubara, ECML’07)
58:15	Refinement vs. calibration plot (1)
58:34	Refinement vs. calibration plot (2)
59:27	Refinement vs. calibration plot (3)
59:53	Ranking and probability estimation — summary
60:54	Concluding remarks
61:39	Some open questions

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.