Initial approximations based on approximate methods are refined using much higher levels of theory.
The key ingredient of our technique is a characterization for one-sided approximations based on orthogonal polynomials.
To estimate the probability of Gambler's Ruin, one can use approximations based on moments (Ethier and Khoshnevisan 2002; Canjar 2007; Hürlimann 2005).
It is shown that, for the rectifying section, by using binary approximations based on the least volatile component, one can obtain reasonable estimates (within a factor of two) for the number of stages.
The derivatives computed by automatic differentiation are compared with approximations based on divided differences.
The data have been analyzed in terms of theoretical approximations based on the Zimm model.
This paper extends previous works by developing plant approximations based on the single and double integrator models with dead time.
Design optimization using approximations based on feed-forward back-propagation neural network is the topic of much recent research.
This proof is obtained by using heat semigroups approximations based on the truncation of Brownian Chen series.
Related problems concerning the computational aspects of iterative calculations and best approximations based on fixed point theory have been also investigated.
All the results also mean that our proposed approximations based on our firsthand observations are reasonable and effective for fast N-best rescoring.
