Cum să cumpăr
Livrare
Consilier de cumpărături
Models for Calculating Confidence Intervals for Neural Networks
engleză
173 b
30 de zile pentru retur bunuri
Ar putea de asemenea, să te intereseze
/
Carte broșată
410
lei
This books provides the methodology of analyzing§existing models to calculate confidence intervals on§the results of neural networks. The three techniques§for determining confidence intervals determination§were the non-linear regression, the bootstrapping§estimation, and the maximum likelihood estimation.§The neural network used the backpropagation algorithm§with an input layer, one hidden layer and an output§layer with one unit. The hidden layer had a logistic§or binary sigmoidal activation function and the§output layer had a linear activation function. These§techniques were tested on various data sets with and§without additional noise. The ranges and standard§deviations of the coverage probabilities over 15§simulations for the three techniques were computed.
Informații despre carte
engleză
