Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/jrnold/stan_headers
Reusable Stan code in CPP macros
https://github.com/jrnold/stan_headers
Last synced: 29 days ago
JSON representation
Reusable Stan code in CPP macros
- Host: GitHub
- URL: https://github.com/jrnold/stan_headers
- Owner: jrnold
- Created: 2013-07-21T20:18:15.000Z (over 11 years ago)
- Default Branch: master
- Last Pushed: 2013-08-21T12:03:37.000Z (about 11 years ago)
- Last Synced: 2024-05-09T22:14:44.464Z (6 months ago)
- Size: 172 KB
- Stars: 1
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Stan Macros Collection
*Stan* (as of version 1.3.0) does not support user-defined functions.
One way to approximate user-defined functions and create reusable code
is by defining macros and
This project contains
# Using
Using CPP will add another step to your workflow. The C pre-processor
must be run on the file in order to expand the macros and generate
valid *Stan* code.
I use the following naming conventions for files:
- `*.stan` Stan model
- `*.stanpp` Stan model containing macros
- `*.stanh` File containing macro definitions for Stan code.
Suppose that `foo.stanpp` is a *Stan* model containing macros.
To generate a valid *Stan* model file named `foo.stan`, run it through the C pre-processor,
```console
cpp -I/path/to/headers -w -undef -Wundef -nostdinc -std=c11 -p foo.stanpp foo.stan
```
A makefile rule for this looks like this,
```makefile
STAN_HEADERS = /path/to/headers
%.stan : %.stanpp
cpp -I$(STAN_HEADERS) -w -undef -Wundef -nostdinc -std=c11 -P $< $@
```
# Documentation
All macros names are ALL CAPS with a leading `_`; this means they are invalid *Stan* variable or function names, making it easier to catch errors.
## _JEFFREYS
`_JEFFREYS(x)` : The Jeffreys prior for variance or scale. This is an improper distribution. Add $1 / x$ to the log-posterior.
## _MAKE_SYMMETRIC
`_MAKE_SYMMETRIC(x)` : If `x` is a matrix, ensure that it is symmeteric with the transformation `0.5 * (x + x')`.
## _TRIANGLE_LOG
`_TRIANGLE_LOG(y)`. The density of the triangle log. `log1m(fabs(y))`.
This assumes that `y` is defined as a bounded variable, e.g.
`real y;`.
## _KALMANF_SEQ
Log-likelihood of a dynamic linear model calculated using Kalman
filter using sequential processing of observations and not allowing
for any missing values.
This calculates the log-likelihood of
$$
y_t \sim N(F'_t \theta_t, V_t)
\theta_t \sim N(G \theta_t, W_t)
$$
```
_KALMANF_SEQ(r, n, y, F, G, V, W, m0, C0)
```
- `r`: `int` Number of observations.
- `n`: `int` Number of latent states
- `y`: `matrix` of data. Rows are variables, columns are observations.
- `F`: `matrix` see equation
- `G`: `matrix` see equation
- `V`: `vector` the diagonal of $V_t$. See equation
- `W`: `matrix` see equation
If `F` (`G`, `V`, or `W`) is an array of matrices, use the following
syntax `F[KALMANF_i]` where `F` is the name of variable.
## _KALMANF
Log-likelihood of a dynamic linear model calculated using Kalman
filter the Kalman filter and not allowing for any missing values.
This calculates the log-likelihood of
$$
y_t \sim N(F'_t \theta_t, V_t)
\theta_t \sim N(G \theta_t, W_t)
$$
```
_KALMANF(r, n, y, F, G, V, W, m0, C0)
```
- `r`: `int` Number of observations.
- `n`: `int` Number of latent states
- `y`: `matrix` of data. Rows are variables, columns are observations.
- `F`: `matrix` see equation
- `G`: `matrix` see equation
- `V`: `matrix` see equation
- `W`: `matrix` see equation
If `F` (`G`, `V`, or `W`) is an array of matrices, use the following
syntax `F[KALMANF_i]` where `F` is the name of variable.
## _LP
`_LP(x)` is equivalent to
```stan
lp__ <- lp__ + x;
```