Re: DM: Request For DM Info

[Ken Collier]

Thomas, 

One problem you will inevitably face is  that  an  evaluation  of
tools  is  almost  immediately  obsolete.  In the Center for Data
Insight (http://insight.cse.nau.edu) we have most of the  leading
commercial  data  mining  tools. All of the tool vendors are very
competitive and are  extremely  active  in  producing  newer  and
better  versions  of  their  tools. These revisions are generally
substantial and include: greater variety of algorithms,  database
connectivity,    client/server   architectures,   improved   data
preprocessing support, etc.

At the CDI we set out to publish a tool  evaluation  and  quickly
realized  that  it  would  be  a  matter  of months before it was
obsolete beyond usefulness. So, we developed  a  tool  evaluation
methodology  based  on our experiences with a variety of business
clients  with  a  variety  of  needs  and  environments.  In  our
experience  the best tool for company X may not meet the needs of
company  Y.   So  it  is  difficult  to  evaluate  tools  without
considering  their  purpose  or  who  will  be  using them. Under
certain circumstances Unica's Model 1 is the best tool and  under
other  circumstances  Thinking  Machine's  Darwin  is the tool of
choice. Our methodology is a scoring process that is based on the
specifics of the end-user's environment.

The CDI exists to  help  companies  understand  data  mining  and
decide  on the best tool for their needs. We use our framework at
the end of each client visit to help them quantify their opinions
about  the  tools after spending time with each one. The feedback
we have recieved from these clients is extremely positive.

Our evaluation method will be published in  the  upcoming  Hawaii
International  Conference on System Sciences (HICSS) Proceedings,
but you can look at a preliminary copy on our website if you  are
interested.

Good luck with your dissertation. I would be  happy  to  talk  DM
techniques with you on another thread.

---
Ken Collier, Ph.D.                      Associate Professor of 
Computer Science
Box 15600                               e-mail: Ken.Collier@nau.edu
College of Engineering and Technology   Phone: 520.523.5412 
Northern Arizona University             Fax: 520.523.2300
Flagstaff, Arizona  86011-1560          WWW: 
http://www.cse.nau.edu/~kwc

Thomas Bradley writes:
> 
> I am a student at the University of Paisley, Scotland, UK.  I am 
> currently undertaking an Honours year project in Data Mining as 
>part of 
> my BA (Hons) Business Information Technology degree.
> 
> The title of my dissertation is "Data Mining; An Evaluation of 
>Tools and 
> Techniques". 
> 
> I would be most grateful if anyone could help me with information 
>on the 
> subject.  For example, history, background, methodologies, tools 
>etc.
> 
> Thanks in advance,
> 
> Thomas A Bradley