Tech Report CS-92-47
On Spline Approximations for Bayesian Computations
Eugene Santos Jr.
Probabilistic reasoning suffers from NP-hard implementations. In particular, the amount of probabilistic information necessary to the computations is often overwhelming. For example, the size of conditional probability tables in Bayesian networks has long been a limiting factor in the general use of these networks.
We present a new approach for manipulating the probabilistic information given. This approach avoids being overwhelmed by essentially compressing the information. Furthermore, unlike existing techniques, we do not need to explicitly access all the data in order to perform our computations. We have achieved this by transforming the problem into linear constraint satisfaction and using approximating functions.