 |
|
|
|
|
 |
 |
|
Discriminant Analysis and Multidimensional Scaling
|
|
|
|
|
|
| Organization: | Charles University of Prague |
|
|
|
|
|
|
2005 Wolfram Technology Conference
|
|
|
|
|
|
Champaign IL
|
|
|
|
|
|
Implementation of two methods of multivariate statistical analysis will be presented.
In the first part on discriminant analysis (DA), the estimation of the statistical characteristics needed for DA will be based on the standard Mathematica statistical package. Afterwards, the algorithms developed for construction of discriminant score functions for decision making will be given. Bayesian methods, maximum likelihood, linear and quadratic DA will be considered.
An algorithm for the classical solution to the multidimensional scaling (MDS) problem will be given in the second part. A configuration of points in Euclidean space using the information on distances between objects only can be constructed.
|
|
|
|
|
|
|
|
|
 |
|
|
| TechConf2005_presentationJH2.nb (418.3 KB) - Mathematica Notebook |
|
|
For the newest resources, visit Wolfram Repositories and Archives »
