gof.optimize: gof.optimize
In snopgof: Single-observation NOn-Parametric Goodness Of Fit

Description Usage Arguments Details Value Author(s) References See Also Examples

View source: R/gof.optimize.R

An optional function which can be used to optimize the weighting matrix against some simulated alternative for a specified type one error level.

1 2	gof.optimize(null, alternate, typeOneError = 0.05, weights = NULL, verbose=FALSE, sannIterations=1000, permIterations=1023)

`null`	A gof.preprocess object containing the preprocessed simulated draws from the null distribution.
`alternate`	A matrix of draws from the desired alternate distribution. Power will be maximized to detect draws from this distribution at the specified type one error level. The coordinates correspond to columns so that each draw corresponds to a row.
`typeOneError`	The desired level of type one error, i.e., at which level will you decide to reject the null hypothesis in favor of the alternative.
`weights`	An initial value for the optimization problem.
`verbose`	Whether to report more or less incremental information while the annealing is optimizing.
`sannIterations`	How many iterations of simulated annealing to attempt. If zero, only the permutations will be considered
`permIterations`	How many permutations to consider, at a maximum. Meaningful thresholds are necessarily in the set 1 3 7 15 31 63 127 255 511 1023 2047 4095 8191 16383 32767 65535 131071 262143 524287 1048575. The default is set to exhaust ten coordinates. When the threshold is hit, the most inclusive permutations are included first.

The optim function is used with the simulated annealing option ("SANN") to optimize power for the given type one error. The results may or may not be satistfactory, and may require some tweaking. This function will try all combinations of diagonal elements as weights in an attempt to find a good solution.

A weighting matrix which may be used as input to gof

Josh Lospinoso

http://stats.ox.ac.uk/~lospinos

snopgof