Maybe the paper
"Sequential Update of Bayesian Network Structures," Nir Friedman and Moises
Goldszmidt In Proceedings of CUAI-97.
can be useful
It can be found online at www.robotics.stanford.edu/~moises
M
Nong Ye wrote:
> I am interested in work and literature on the incremental learning of
> Bayesian networks.
>
> Also, any literature on the incremental learning of artificial neural
> networks?
>
> Let's say, I have trained a Bayesian network or artificial neural network
> usig a set of data. Once I used this set of data, I discarded it because
> I do not have space to store it. Now a new set of data becomes available
> for the trained Bayesian network or the trained artificial neural network
> to do further training on the new set of data. How do I perform more
> training
> on the trained Bayesian network or the artificial neural network to ensure
> that the twice-trained Bayesian network or artificial neural network
> fit to both the old set and the new set of data, given I no longer have
> the old set of data?
>
> Nong Ye
> Arizona State University
>
