apop_exponential.c File Reference

Functions

apop_rank_settings * apop_rank_settings_init (apop_rank_settings in)
apop_rank_settings * apop_rank_settings_alloc (void *ignoreme)
void apop_rank_settings_free (apop_rank_settings *in)
void * apop_rank_settings_copy (apop_rank_settings *in)

Variables

apop_model apop_exponential

Detailed Description

The Exponential distribution.

Function Documentation

apop_rank_settings* apop_rank_settings_init ( apop_rank_settings in )

If this settings group is present, models that can take rank data will read the input data as such. Allocation is thus very simple, e.g.

  Apop_model_group_add(your_model, apop_rank);

Variable Documentation

apop_model apop_exponential

The Exponential distribution. A one-parameter likelihood fn.

Ignores the matrix structure of the input data, so send in a 1 x N, an N x 1, or an N x M.

$Z(\mu,k) = \sum_k 1/\mu e^{-k/\mu}$
$ln Z(\mu,k) = \sum_k -\ln(\mu) - k/\mu$
$dln Z(\mu,k)/d\mu = \sum_k -1/\mu + k/(\mu^2)$

Some write the function as: $Z(C,k) dx = \ln C C^{-k}.$ If you prefer this form, just convert your parameter via $\mu = {1\over \ln C}$ (and convert back from the parameters this function gives you via $C=\exp(1/\mu)$ .

To specify that you have frequency or ranking data, use