Recebemos um mês depois, aproximadamente, a seguinte resposta de Guastello:
"Wal,
It looks as though you've tried a lot of things already.
Two suggestions at this stage. One is to find a variable in your data set that distinguishes the results from the other analyses, e.g. cases with high entropy versus cases with low entropy.
You'd be essentially looking for control variables that change the dynamics.
If you have many dynamical properties to sift through, try making a new data set, N = 200, with all the dynamical variables and factor analyze them.
That should reduce the mischief level a bit.
-Stephen"
quinta-feira, 19 de março de 2009
Assinar:
Postar comentários (Atom)
Nenhum comentário:
Postar um comentário