Transcript Presentation by Dragan Gamberger

WP4
Zagreb, 21.june 2007
WP4
Knowledge representation, discovery, and management
T4.1 Data warehouse for knowledge discovery
T4.2 Data understanding and preparation
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation
WP4
Knowledge representation, discovery, and management
ended 1.may
T4.1 Data warehouse for knowledge discovery
T4.2 Data understanding and preparation
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation
to end on 1. august
♦ MS4.1
!!!
WP4
Knowledge representation, discovery, and management
ended 1.may
T4.1 Data warehouse for knowledge discovery
T4.2 Data understanding and preparation
till 1. september
till M33
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation
to end on 1. august
♦ MS4.1
!!!
WP4
Knowledge representation, discovery, and management
T4.1 Data warehouse for knowledge discovery
D21
T4.2 Data understanding and preparation
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation
to end on 1. august
♦ MS4.1
!!!
WP4
Knowledge representation, discovery, and management
T4.1 Data warehouse for knowledge discovery
D21
T4.2 Data understanding and preparation
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation
D22
♦ MS4.1
WP4
Knowledge representation, discovery, and management
T4.1 Data warehouse for knowledge discovery
D29 month 24
T4.2 Data understanding and preparation
MS4.2
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation
♦ MS4.1
WP4
Knowledge representation, discovery, and management
T4.1 Data warehouse for knowledge discovery
T4.2 Data understanding and preparation
D21
D29
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation D22
WP4
Knowledge representation, discovery, and management
T4.1 Data warehouse for knowledge discovery
T4.2 Data understanding and preparation
D21
D29
T4.3 Implementation of knowledge discovery processes
T4.4 Ontologies and knowledge representation D22
+ integration of knowledge representation with DSS
WP3
WP5
WP4
HF Middleware
D
S
S
Knowledge Base
During the review meeting in Brussels pointed out as
very relevant and very difficult problem
What has been done in the period from the
last meeting:
T4.1
- Functional specification of the data transformation from
patient data defined by the 'Case report form' into the
complete set of attributes prepared for knowledge discovery
has been done.
- Open source DW software package Pentaho has been
extensively studied and its application tested for some multipatient managerial tasks and single patients decision support
purposes.
- Work on D21.
T4.2
- Retrospective data from UNICZ have been transformed to
the standard Weka (‘arff’) and ‘illm’ form. Data have been
cleaned manually by elimination of obvious outliers and
sparsely populated attributes.
- Retrospective ANMCO data set has been transformed to
the Weka format.
- D21 on data preparation has been started.
T4.3
- Experiments with contrast set mining and Kernel based Support
Vector Machine (?) algorithms have started in parallel on
available retrospective data.
-ANMCO seems as a good simulation of data that will be
collected by the platform (except continuous data monitoring).
- We have some results with UNICZ and ANMCO data, but the results
are not very optimistic.
T4.4
- HF ontology has been upgraded.
- Meeting in Milan with UNIMIB and AUXOL ->
procedural knowledge related to HF has been prepared in the
form of 10 sets of rules.
- In cooperation with JUMC some of the rule sets have been
discussed and improved.
- Additionally, for some common HF related decision problems
medical plans have been developed.
ontology + rules + medical plans are on line on
http://lis.irb.hr/heartfaid/
T4.4 continued
encountered problems:
- developed HF ontology is large but not appropriate for
reasoning (taxonomy) !
- Open issue: how to integrate ontology (descriptive knowledge)
and rules (procedural knowledge) ?
possible solution:
development of the procedural ontology + development of the
ontology interpreter for DSS interface
we have devoted a lot of work to this idea!
WP4 presentation plan:
- Knowledge representation for HF domain (D22) - Alan Jović
- Demo on procedural ontology and ontology reasoning for
DSS support - Marin Prcela
- Data warehousing on HF platform (D21) - Tom Šmuc
- Discussion on:
a) D21
b) D22
c) KB – DSS interface