The Complete SPRLIB & ANNLIB

fisherq: - determine the coefficients for the quadratic decision function

SYNOPSIS

int fisherq

(Ka, Kb, Ma, Mb, c, n, W2, W1, b)

ARGUMENTS

`double`	`Kainv`**	The covariance matrix for class A, a `matrix` of size `[1..n][1..n].`
`double`	`Kbinv`**	The covariance matrix for class B, a `matrix` of size `[1..n][1..n].`
`double`	*`Ma`**	The mean vector for class A, a `vector` of size `[1..n].`
`double`	*`Mb`**	The mean vector for class B, a `vector` of size `[1..n].`
`double`	`c`	The quotient of the apriori probabilities.
`int`	`n`	The dimensionality of the feature space.
`double`	`W2`**	The quadratic coefficient, a `matrix` of size `[1..n][1..n].`
`double`	*`W1`**	The linear coefficient, a `vector` of size `[1..n].`
`double`	*`b`**	The constant term.

RETURNS

The function returns TRUE if an error was detected, and FALSE otherwise. If no error occured the variables W2, W1, and b contain valid values for the coefficients.

FUNCTION

The function computes the coefficients needed for using the quadratic decision function. This quadratic function is given by x^T*W2*x + W1*x + b. In order to compute these coefficients the covariances and means of both classesare required. Furthermore the quotient c = P(a)/P(b) is needed.

NOTE

The variables b, W1, Ma and Mb are in vector format, Ka and Kb are in matrix format.

The Complete SPRLIB & ANNLIB

`fisherq`

- determine the coefficients for the quadratic decision function

SYNOPSIS

ARGUMENTS

RETURNS

FUNCTION

NOTE

SEE ALSO