Just to remind everyone about Paul's talk this coming Monday!
Thierry C.
--------------------------------------------------------------------------------
Data Mining Applications - Open Lectures Series
(Joint meeting of the OR Society Health and Social Services Study Group;
Hosted by the University of Westminster School of Informatics)
--------------------------------------------------------------------------------
MONDAY 11 February 2008, 16:45-18:00 (talk starts at 17:00)
115 New Cavendish Street, Room C1.04, University of Westminster, London
http://www.streetmap.co.uk/streetmap.dll?G2M?X=529158&Y=181858&A=Y&Z=1
--------------------------------------------------------------------------------
TreeFit: the development and application of a powerful decision tree data
mining tool
Prof Paul Harper, Cardiff University
--------------------------------------------------------------------------------
Abstract
This talk will present TreeFit, a powerful and user-friendly decision tree
software that has been primarily developed for use by the NHS Information
Centre (IC). The IC is using TreeFit to assist in the revision of
Healthcare Resource Groups (HRGs). HRGs are standard groupings of clinically
similar treatments, which use common levels of healthcare resource. They may
be considered as "units of currency" within the health service, allowing for
costings across services, and will support the UK Department of Health's
policy of Payment by Results.
Decision trees are data mining tools which are used to represent rules
underlying data. Decision trees are hierarchical, sequential classification
structures that recursively partition the set of observations (data). The
interest of our research was is to create a tool that resulted in more
valuable decision trees by enhancing existing theory. In particular, we are
considering issues relating to scalability, hybrid trees and the development
of "fuzzy trees". We will present the latest version of TreeFit which
includes the following functionality:
1. The ability to handle large data sets.
2. The capacity to cope with data that contains several distinct values
of categorical data.
3. Provision of validation of the tree model.
4. Possibility to move data across terminal nodes.
5. Description of the nodes path.
6. Visualisation of the data present in a training model.
7. Support to missing values in the data.
8. Allowing the user to interact with the model, choosing the split
variable, split values, grow and prune subtrees.
--------------------------------------------------------------------------------
Speaker's biography
Paul Harper is Professor of Operational Research at Cardiff University and
holds the position of WIMCS (Welsh Institute for Mathematical and
Computational Sciences) Research Chair in OR. His research interests include
healthcare modelling and he has published widely on modelling topics ranging
from health service planning (for example hospital capacity models,
workforce redesign and facility location problems) to public health policy
(for example strategies for screening breast cancer, diabetes and
chlamydia). Professor Harper has worked with a number of healthcare
organisations and has received research funding from various sources
including EPSRC, NHS Information Centre, ESRC and The Nuffield Foundation.
He was awarded the 2006 Goodeve medal by the UK OR Society for his
co-authored paper on targeted screening strategies for Chlamydia (deemed the
best published paper in the Journal of the Operational Research Society in
that year).
--------------------------------------------------------------------------------
For more information contact Prof Thierry Chaussalet, School of Informatics,
Department of Information Systems and Computing, 115 New Cavendish St,
London; Tel: 020 7911 5000 ext 3540; [log in to unmask]
-----------
Prof T.J. Chaussalet
School of Informatics
Department of Information Systems and Computing
University of Westminster
115 New Cavendish Street
London W1W 6UW
Tel: +44(0)207 911 5000
Email: [log in to unmask]
http://www.healthcareinformatics.org.uk
-----------
PLEASE READ: This e-mail and its attachments are intended for the above
named only and may be confidential. If they have come to you in error you
must not copy or show them to anyone, nor should you take any action based
on them, other than to notify the error by replying to the sender.
Any views expressed in this message are those of the individual sender,
except where the sender specifically states otherwise.
--
The University of Westminster is a charity and a company limited by
guarantee. Registration number: 977818 England. Registered Office:
309 Regent Street, London W1B 2UW, UK.
|