Re: DM: KnowledgeSEEKER

[Ken Collier]

Tjen-Sien Lim writes:
> 
> I'd be interested to hear any opinion/comment from people who have
> used KnowledgeSEEKER. For what sort of applications will
> KnowledgeSEEKER be useful? Any applied papers that mention KS?
> Thanks.

We have been using KnowledgeSEEKER  along  with  many  other  KDD
tools  in the Center for Data Insight for the past year. Recently
we have replaced that with the newer KnowledgeStudio from Angoss.

KnowledgeSEEKER is a very robust tool  that  handles  data  in  a
variety  of  formats.  It produces CARTŪ  and CHAID trees in a very
expressive and comprehensible manner. It supports  the  automatic
growing  of  trees  or allows the user to guide tree growth along
interesting paths. I am most impressed with KS' ability to handle
data of questionable quality. Missing values or inconsistent data
types don't create problems for KS as  they  do  for  some  other
tools.  Also,  KS  handles larger data sets without a hitch where
other tools take days to build  models.  The  interface  is  very
intuitive  and  the tool has a very nice learning curve. We often
use KS as a  first  tool  for  understanding  our  clients'  data
better.

You could use KS anywhere you want to analyze  the  relationships
between  independent  variables  and  a  target  variable  (i.e.,
directed knowledge discovery or hypothesis testing).  The  target
variable  can  either  be  categorical or continuous. I would not
consider  this  an  appropriate  tool  for  undirected  knowledge
discovery   since   that  type  of  application  is  generally  a
clustering activity.

KnowledgeStudio implements a  wider  variety  of  algorithms  for
analysis. So, it has all the great things about SEEKER plus more.
Angoss is definitely a  player  and  sets  a  standard  for  tool
robustness. Hope this helps.

---
Ken Collier, Ph.D.
Technical Director, Center for Data Insight     Email: 
Ken.Collier@nau.edu
College of Engineering and Technology           Phone: 520.523.5412 
Northern Arizona University                     Fax: 520.523.2300
Flagstaff, Arizona  86011-1560                  
http://insight.cse.nau.edu