err_abs

Purpose

Computes the absolute error.

Synopsis

[absAvg, absInd] = err_abs(likpar,y,mX,sigX2)

Description

[absAvg] = err_abs(likpar,y,mX,sigX2) - returns the sum of absolute errors.

The additional second output argument returns the individual absolute errors: [absAvg, absInd] = err_abs(likpar,y,mX,sigX2)

The input argument likpar is not used.

Parameters:

likpar

- the Gaussian Process data structure.

y

- desired output.

mX

- mean of the GP marginal at the new input.

absAvg

- the average absolute error.

absInd

- the vector of individual absolute errors.

See Also

ogp, ogptrain, demogp_reg, c_reg_lapl, c_reg_exp

Copyright (c) Lehel Csató (2001-2004)