Veaux Evelyn Rosengarten's scientific contributions

Publication (1)

Article
Neural networks can be viewed as nonlinear models, where the weights are parameters to be estimated. In general two parameter estimation methods are used: nonlinear regression, corresponding to the standard backpropagation algorithm, and Bayesian estimation, in which the model parameters are considered as being random variables drawn from a prior d...

Citations

... Therefore, even though the stochastic gradient descent approach has proven to obtain good results while minimizing computational costs, research on neural interval models has also focused on evolutionary computing solutions, such as Genetic Algorithms [67] and Particle Swarm Optimization [68] [69] . Additionally, some niche cases also exist where the prediction interval construction method does not allow for any of these solutions, and instead requires a specific training methodology, such as Bayesian approaches [73] , which rely on Markov Chain Monte Carlo sampling algorithms. ...