Subgroup Discovery: Recent Biomedical Applications

author: Nada Lavrač, Jožef Stefan Institute

Description

This talk presents recent advances in data mining, focusing on subgroup discovery and the ways to use subgroup discovery to generate actionable knowledge for decision support. Actionable knowledge is explicit symbolic knowledge, typically presented in the form of rules, that allow the decision maker to recognize some important relations needed to perform an appropriate action, such as planning a population screening campaign aimed at detecting individuals with high disease risk. Different subgroup discovery approaches are outlined, illustrated with case studies from medicine and functional genomics.

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Subgroup Discovery: Recent Biomedical Applications
0:21	Talk outline
1:00	Data Mining in a Nutshell (1)
1:34	Data Mining in a Nutshell (2)
2:01	Data mining example Input: Contact lens data
2:24	Output: Decision tree for contact lens prescription
2:27	Output: Classification/prediction rules for contact lens prescription
2:30	Task reformulation: Concept learning problem (positive vs. negative examples of Target class)
3:16	Classification versus Subgroup Discovery
3:25	Task reformulation: Concept learning problem (positive vs. negative examples of Target class)
3:31	Classification versus Subgroup Discovery
5:31	Classification versus Subgroup Discovery
6:29	Talk outline
6:36	Subgroup discovery in a nutshell
7:26	CHD Risk Group Discovery Task
8:42	Subgroup discovery in the CHD application
12:41	Induced subgroups and their statistical characterization
12:59	Statistical characterization of expert selected subgroups
13:00	Statistical characterization of subgroups
14:41	Propositional subgroup discovery algorithms
15:35	Characteristics of Subgroup Discovery Algorithms
17:58	Weighted covering algorithm for rule set construction (1)
18:16	Weighted covering algorithm for rule set construction (2)
22:10	Weighted covering algorithm for rule set construction (3)
22:15	Weighted covering algorithm for rule set construction (2)
22:37	Weighted covering algorithm for rule set construction (3)
22:53	Weighted covering algorithm for rule set construction (2)
23:09	Weighted covering algorithm for rule set construction (3)
23:17	Subgroup visualization
24:21	Statistical characterization of expert selected subgroups
25:41	Subgroup discovery lessons learned
26:55	Talk outline
27:03	Relational Data Mining (Inductive Logic Programming) in a Nutshell
27:52	Relational Data Mining (ILP)
29:36	RSD: Upgrading CN2-SD to Relational Subgroup Discovery
29:37	Propositionalization in a nutshell
29:45	Relational Data Mining (ILP)
30:30	RSD: Upgrading CN2-SD to Relational Subgroup Discovery
30:54	Propositionalization in a nutshell
34:17	Propositionalization in relational data mining
37:14	Propositionalization in a nutshell
37:47	Propositionalization in relational data mining
37:57	Propositionalization in a nutshell
38:20	Propositionalization in relational data mining
39:54	Relational subgroup discovery
40:54	Propositionalization in relational data mining
41:11	Relational subgroup discovery
41:39	RSD Lessons learned
43:18	Talk outline
43:34	DNA microarray data analysis
44:21	Gene Expression Data: data mining format
45:13	Standard approach: High-Dimensional Classification Models
45:47	High-Dimensional Classification Models (cont’d)
46:09	Subgroup discovery in DNA microarray data analysis: Functional genomics domains
47:18	Subgroup discovery in microarray data analysis
47:58	Subgroup discovery in microarray data analysis: Extert’s comments
48:27	Subgroup discovery in microarray data analysis
48:32	Subgroup discovery in microarray data analysis: Extert’s comments
49:00	Propositional subgroup discovery: Accuracy-Interpretability trade off
49:33	Talk outline
50:20	Accuracy-Interpretability trade off
52:41	Actual approach approach to Learning 1: Identifying sets of differentially expressed genes in data preprocessing
53:01	Identifying diffferentially expressed genes (1)
53:36	Identifying diffferentially expressed genes (2)
54:15	Ranking of differentially expressed genes
54:47	Statistical Significance Meets Biological Relevance: Motivation for relational feature construction
55:28	Relational Subgroup Discovery
56:13	Gene Ontology (GO)
57:29	Gene Ontology (2)
57:40	Multi-Relational representation
58:01	Encoding as relational background knowledge
58:59	RSD First order feature construction
60:59	Propositionalization
61:49	Propositional learning: subgroup discovery
62:09	Subgroup Discovery (1)
62:15	Subgroup Discovery (2)
62:27	Subgroup Discovery (3)
62:28	Summary: The RSD approach
63:10	Experiments
63:59	Results – Discovered subgroup descriptions
64:25	Results – Discovered subgroup descriptions (2)
64:28	Results – Clear effect of using background knowledge and weigths in learning
64:54	Related and Recent work
65:45	Summary of the presented RSD approach
66:44	Summary of recent work
66:47	Future work:Towards service-oriented knowledge technologies for information fusion

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.