[Jtreeview-users] consistent alternatives of triangle matirx with average linkage method

[Thorsten O. <tho...@ly...>]

Hello,

we would like to work with cluster 3.0 and treeview 1.1.0  for generating=20
consistent alternatives of an evaluation of a triangle matrix. The rows=20
and columns have got the same signification/variables.
My question is how to get a choosen number of alternatives with the=20
average linkage method and if i have to modify something in the cluster or=
=20
the treeview program.
Regarding the evaluation of the connections/relation there has to be=20
processed a clustering process looking for connection each other=20
supporting. For instance, if variable "a" is connected with "b" by a 5 (5=20
means very good support), but "a" to "c" by a 3 (more or less/or=20
independent), and "b" to "c" by 5 the process has to figure out that "a"=20
is supporting "c" through its very support with "b". The only difference=20
should be that the 2nd degree connection has another weight, means that=20
there is considered less value than for a direct connection.

We do need this kind process concluding a number of alternatives to be=20
choosen. The values we=B4re using are 1 to 5, but they may change. It=20
doesn=B4t matter!
1 =3D completly inconsistent influence
2 =3D probably inconsistent influence
3 =3D not influencing each other / independent
4 =3D probably consistent influence
5 =3D completly consistent influence

It would be great if i might get in touch with somebody of capacity=20
personally.
Thank a lot for your effort applied in advance!

Best regards!

Thorsten


Thorsten Ortjohann
Consultant Business Process Management
Lynx-Consulting AG
Johanniskirchplatz 6
D-33615 Bielefeld
Fon: +49 521 5247-0
Fax: +49 521 5247-250
Mobil: +49 175 22 85 490


Company and Management Headquarters:
Lynx-Consulting AG, Johanniskirchplatz 6, D-33615 Bielefeld
Fon: +49 521 5247-0, Fax: +49 521 5247-250, www.lynx.de

Court Registration: Amtsgericht Bielefeld HRB 35946
Members of the Management Board authorised to represent: Peter H=FCsener
Chairman of the Supervisory Board: Hans-Joachim Rosowski

---------------------------------------------------------------------------=
-------------------------
This e-mail may contain trade secrets or privileged, undisclosed, or otherw=
ise confidential information. If you have received this e-mail in error, yo=
u are hereby notified that any review, copying, or distribution of it is st=
rictly prohibited. Please inform us immediately and destroy the original tr=
ansmittal. Thank you for your cooperation.=20
---------------------------------------------------------------------------=
-------------------------

=20