macroeco.compare.AIC_compare

macroeco.compare.AIC_compare(aic_list)

Calculates delta AIC and AIC weights from a list of AIC values

Parameters:

aic_list : iterable

AIC values from set of candidat models

Returns:

tuple :

First element contains the delta AIC values, second element contains the relative AIC weights.

Notes

AIC weights can be interpreted as the probability that a given model is the best model in the set.

Examples

>>> # Generate random data
>>> rand_samp = md.nbinom_ztrunc.rvs(20, 0.5, size=100)

>>> # Fit Zero-truncated NBD (Full model)
>>> mle_nbd = md.nbinom_ztrunc.fit_mle(rand_samp)

>>> # Fit a logseries (limiting case of Zero-truncated NBD, reduced model)
>>> mle_logser = md.logser.fit_mle(rand_samp)

>>> # Get AIC for ztrunc_nbinom
>>> nbd_aic = comp.AIC(rand_samp, md.nbinom_ztrunc(*mle_nbd))

>>> # Get AIC for logser
>>> logser_aic = comp.AIC(rand_samp, md.logser(*mle_logser))

>>> # Make AIC list and get weights
>>> aic_list = [nbd_aic, logser_aic]
>>> comp.AIC_compare(aic_list)
(array([  0.        ,  19.11806518]),
array([  9.99929444e-01,   7.05560486e-05]))

>>> # Zero-truncated NBD is a far superior model based on AIC weights