H-methods give new mathematics within several fields of applied mathematics.
|Double Weighing Schemes|
|Multi-Way Data Analysis|
|Polynomials Surfaces in low dimensions|
In path modelling the methods of regression analysis is extended to
Traditional methods typically work with some measures of correlations between data blocks in a path.
The present methods provide with latent structures for each data block based on the path
that has been specified. See a review of the methods.
H-methods have been extended to a directed network of data blocks. Again the
methods and techniques
Regression analysis have been extended to this case. See a review of methods.
H-methods have been applied to causality analysis. The basic ideas are to
take the loading matrix and work further
with different methods. An example is to replace the loading values by values that are significantly different from zero.
For a short review see an introduction.
There have been developed double weighing schemes to analyse data. It means
that weights are defined for both
variables and samples. This provides with new methods to analyse tables. The methods extend to systems like
'biological systems' consisting of interacting data blocks. See a review of the basic principles.
H-methods have been extended to multi-way data analysis. An important aspect
of the methods is the
definition of inverse in different directions. See a short review.
H-methods have been extended to low rank non-linear analysis. The methods are
of two type. The first type is
extension of linear latent structure to a latent structure as multivariate polynomials in score variables. The other type
is to use low rank solutions in establishing non-linear solutions.
Back to Main Page.