| |
|
Science Forum Index » Statistics - Education Forum » MDS
Page 1 of 1
|
| Author |
Message |
| zagatov |
 Posted: Tue Jan 02, 2007 3:58 pm |
|
|
|
Guest
|
Hi
I studying with MdS (MultiDimensional Scaling)
If anyone interested in , appriciate
I try to reconstruct Alscal and other algorithm for obtain same input
/output
same as SPSS .I used Java to construct it.
First step Obtain Proximity or Dissimilarity from variables
Second step Normalization of Matrix
Third step Diagonalization and finding coordinates
4. step Stress of d - d hat
and then iterative step which is solved by Newton-Rhapson or Steepest
Descent
all steps respectively
Householder Tridiagonalization
QR Algorithm
Normalization of the Matrix
Steepest Descent with Stress Value (again first step)
Estimation of f(d)=a+xb is iteratively getting closer to D hat
Am I right ? Do you know anything about right steps out of the
algorithm of SPSS? |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Wed Dec 03, 2008 9:44 pm
|
|
