<< nns_simulbis (X->y) Non-linear calibrations i-Asca-anova >>

fact >> (X->y) Non-linear calibrations > nns_simulter

nns_simulter

prediction errors by a neural network with 1 hidden layer and with bias

Calling sequence

[sse,gra,hes]=nns_simulter(wh,wo,x,y)

Arguments

wh:

coefficients of the hidden neurons + bias

wh is a matrix of dimensions ((q+1) x nh) or a Div structure

nh is the number of hidden neurons

wo:

coefficients of the output neurons + bias

wo is a matrix of dimensions ((nh+1) x no) or a Div structure

no is the number of output neurons

x:

calibration dataset

x is a matrix of dimensions (n x q) or a Div structure

y:

reference values, to be predicted

y is a matrix of dimensions (n x no) or a Div structure

sse:

sum of squares of the errors, for each output

sse is a Div structure

sse.d is a vector of dimensions (1 x no)

gra:

partial derivatives of the sums of square of the errors, with respect to each weight, for each output

gra is a Div structure

gra.d is a matrix of dimensions (nw x no)

nw=(q+1)*nh + (nh+1)*no

hes:

Hessians, approximate second order derivatives of the sums of squares of the errors with respect to the weights, for each output

hes is an hypermatrix of dimensions (nw x nw x no) or a Div structure

Examples

[sse,gra,hes]=nns_simulter(wh,wo,x,y)

Bibliography

Authors


Report an issue
<< nns_simulbis (X->y) Non-linear calibrations i-Asca-anova >>