Royal Statistical Society
Statistical Computing Section Meeting
Wednesday 24 November, 2.00 - 5.00pm (tea at 3.30pm)at
The Royal Statistical Society, 12 Errol St, London EC1
(nearest UG stations, Barbican, Moorgate & Old Street)
DATA MINING: Software and applications
----------------------------------------------
Speakers
TERESA DAVENPORT (SAS LTD)
KATE SWEENEY (Unilever Research)
XIAOHUI LIU, STEVE SWIFT, ALAN TUCKER
(Birkbeck College University of London)
Programme
---------
2.00pm Teresa Davenport Data Mining Tools: SAS Enterprise Miner
(talk, demonstration & discussion)
3.30pm Tea
3.45pm Xiaohui Liu Bridging the gap between applications & tools:
Steve Swift modelling multivariate time series data
Alan Tucker
4.15pm Kate Sweeney Experiences using SAS Enterprise Miner
Synopses
TERESA DAVENPORT (SAS LTD)
Data Mining Tools: SAS Enterprise Miner
Synopsis: The Data Mining process in SAS Enterprise Miner is called SEMMA.
This process will be described, together with the corresponding tools,
and the software will be demonstrated. SEMMA stands for Sample (eg
partitioning into training, validation and test data sets), Explore
(plotting, descriptive statistics, association analysis) Modify (transform,
identify outliers, impute missing values, cluster, analyse with SOMS or
Kohonen networks), Model (decision trees, regression models, neural networks,
user defined models) and Access (comparison of competing predictive models,
chart building etc).
XIAOHUI LIU, STEVE SWIFT, ALAN TUCKER (Birkbeck College, University of London)
Bridging the gap between applications and tools: Modelling multivariate time
series data
How can the gap be bridged between what modern, practical,large scale
applications demand and what conventional methods or tools can offer in the
modelling of multivariate time series (MTS) data? Several real-world
applications will be introduced involving the analysis of MTS data where
existing methods have been found inadequate. Two applications will be
described in depth and recent work using programmed intelligent data analysis
and data mining techniques will be discussed.
KATE SWEENEY (Unilever Reseach)
Experiences using SAS Enterprise Miner
This talk discusses first why Unilever Research are looking at Data Mining and
our approach to the Data Mining trial. Thia will include the choice of software
packages to trial and our broad acceptance criteria. There will also be some
discussion of the sorts of analyses we have been doing with
Enterprise Miner and our general thoughts on SAS EM and
Data Mining to date.
----------------------
Suzanne Evans
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|