Stan

https://github.com/stan-dev/stan
Admin
Stan is a C++ package for Bayesian inference using NUTS, ADVI, and penalized MLE, built on a C++...
Tokens:3,248
Snippets:40
Trust Score:-
License:BSD-3-Clause
Update:1 year ago
Show doc for...
Tokens:
Raw
### APIDOC: Documentation Updates for Operators and Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds documentation for infix operators, the categorical_logit sampling statement, and an example for Cholesky factor with unit diagonal transform, improving clarity and usage guidance.

```APIDOC
add infix operators to manual
categorical_logit sampling statement
Cholesky factor with unit diagonal transform example
```

--------------------------------

### APIDOC: Multivariate Priors for quad_form_diag

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Provides new advice on multivariate priors specifically for the quad_form_diag function, guiding users on best practices for model specification.

```APIDOC
new advice on multivariate priors for quad_form_diag
```

--------------------------------

### APIDOC: CmdStan Integration for Binaries

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Moves bin/print and bin/stanc to CmdStan, indicating a shift in how main executables are generated when compiling models from Stan C++.

```APIDOC
move bin/print and bin/stanc to CmdStan;  no longer generating main
  when compiling model from Stan C++
```

--------------------------------

### APIDOC: Gaussian Process Densities

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces new densities for Gaussian processes: multi_gp_log() and multi_gp(), enabling more complex statistical modeling.

```APIDOC
densities: Gaussian processes multi_gp_log(), multi_gp()
```

--------------------------------

### APIDOC: a^b Syntax for pow(a,b)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds a^b as a syntactic sugar for pow(a,b), providing a more natural and concise way to express exponentiation.

```APIDOC
a^b syntax for pow(a,b)
```

--------------------------------

### Stan CmdStan: Fixed Parameter Sampler

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

A new `fixed_param` sampler was added for use specifically with the generated quantities block, allowing for fixed parameter sampling without MCMC iterations.

```APIDOC
sampler: fixed_param
usage: For generated quantities block
```

--------------------------------

### C++ chainable_alloc Class (v1.3.0)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces a new C++ class, `chainable_alloc`, designed for caching solver results, improving performance and memory management within the Stan C++ backend.

```APIDOC
C++ Class:
  class chainable_alloc {
    // Purpose: Caches solver results for improved performance.
    // Details: Designed for internal use within Stan's C++ backend.
  }
```

--------------------------------

### APIDOC: New Matrix Operations

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds new matrix operations including quad_form_sym() (x' *Sigma * x) and QR decompositions (qr_Q(), qr_R()), expanding the linear algebra capabilities.

```APIDOC
matrix operations: quad_form_sym() (x' *Sigma * x), QR decompositions
  qr_Q(), qr_R()
```

--------------------------------

### APIDOC: Effective Sample Size Documentation Clarification

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Fixes a typo and clarifies the documentation for effective sample size, improving the understanding of diagnostic metrics.

```APIDOC
fixed typo and clarified effective sample size doc
```

--------------------------------

### APIDOC: Zero-Length Inputs for Multivariate Distributions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Allows zero-length inputs for multi_normal and multi_student_t distributions, with an appropriate log probability of 0, enhancing flexibility for edge cases.

```APIDOC
allow zero-length inputs to multi_normal and multi_student_t
  with appropriate log prob (i.e., 0)
```

--------------------------------

### Stan Broadcast/Repetition Functions (v1.3.0)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces functions for creating repeated arrays, matrices, and vectors, simplifying initialization and data manipulation in the Stan modeling language.

```APIDOC
Functions:
  rep_array(value, int dim1, ...): Create an array with repeated value
  rep_matrix(value, int rows, int cols): Create a matrix with repeated value
  rep_row_vector(value, int cols): Create a row vector with repeated value
  rep_vector(value, int rows): Create a column vector with repeated value
```

--------------------------------

### Stan Language: Sort and Rank Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

New `sort` and `rank` functions were added to the Stan language, providing capabilities for ordering and ranking data within models.

```APIDOC
sort_asc(array)
sort_desc(array)
rank(array)
```

--------------------------------

### APIDOC: New Function Signature Tests

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds tests for log_softmax, multiply_lower_tri_self_transpose, and tcrossprod, ensuring the correctness of these new or updated function signatures.

```APIDOC
added tests for log_softmax, multiply_lower_tri_self_transpose, tcrossprod
```

--------------------------------

### APIDOC: pow Syntax for Vector Usage

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Fixes the pow (^) syntax to correctly catch and disallow illegal use with vectors, ensuring proper type handling.

```APIDOC
fix for pow as ^ syntax to catch illegal use of vectors (which
  aren't supported)
```

--------------------------------

### APIDOC: Higher-Order Autodiff

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Completes the implementation of higher-order automatic differentiation, extending support to all probability functions, matrix functions, and matrix operators. Tested up to 3rd order derivatives.

```APIDOC
completed higher-order autodiff (added all probability functions,
  matrix functions, and matrix operators);  tested up to 3rd order
```

--------------------------------

### RStan Exposed Log Probability and Gradient Functions (v1.2.0)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Exposes the model's log probability and gradient functions in RStan, allowing their direct use by external R packages, such as optimizers, for advanced analysis.

```APIDOC
RStan Exposed Functions:
  log_prob(stan_model_object, parameters): Returns the log probability of the model given parameters.
  gradient(stan_model_object, parameters): Returns the gradient of the log probability with respect to parameters.
  // Usage: Intended for integration with external R optimization packages.
```

--------------------------------

### APIDOC: Scalar Operations in Element-Wise and Matrix Division

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Allows the use of scalars in all element-wise operations and enables matrix division for mixing scalars and matrices, increasing flexibility in mathematical expressions.

```APIDOC
added ability to use scalars in all element-wise operations
allow matrix division for mixing scalars and matrices
```

--------------------------------

### APIDOC: main() Removal from Generated Models

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Removes main() from models generated from C++ Stan, making them available only in CmdStan and streamlining the compilation process.

```APIDOC
removed main() from models generated from C++ Stan (they are
  now available only in CmdStan); removed no_main command options
```

--------------------------------

### Stan Language: Negative Binomial Parameterization

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduced an alternative parameterization for the negative binomial distribution, `neg_binomial_2()`, to provide more flexible modeling options within the Stan language.

```APIDOC
neg_binomial_2(y | mu, phi)
```

--------------------------------

### New Stan Probability Distributions (v1.3.0)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces new probability distributions available in the Stan modeling language, expanding the range of statistical models that can be directly specified.

```APIDOC
Distributions:
  exp_mod_normal
  gumbel
  skew_normal
```

--------------------------------

### RStan as.mcmc.list() Method (v1.3.0)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds a method to convert Stan output objects into an 'mcmc.list' object, facilitating compatibility with other R packages for MCMC analysis.

```APIDOC
RStan Method:
  as.mcmc.list(stan_fit_object): Converts a Stan fit object to an mcmc.list object.
```

--------------------------------

### Stan Language: New Mathematical Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Added `distance`, `squared_distance`, and `diag_pre_multiply` to the Stan modeling language for common mathematical operations, enhancing model expressiveness.

```APIDOC
distance(vector1, vector2)
squared_distance(vector1, vector2)
diag_pre_multiply(vector, matrix)
```

--------------------------------

### APIDOC: 1-Indexed Error Messages

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Modifies error messages to be 1-indexed instead of 0-indexed, providing a more user-friendly and consistent experience.

```APIDOC
modified error messages so that they're all 1-indexed instead of 0-indexed
```

--------------------------------

### APIDOC: User-Defined Function Bug Fixes

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Addresses several issues related to user-defined functions, including preventing double compilation, fixing problems with int-only parameters, resolving NaN initialization, and adding checks for variable name conflicts.

```APIDOC
fixed problem with compiling user-defined function twice
fixed problem with int-only parameters for user-defined functions
fixed NaN init problems for user-defined functions
added check that user variable doesn't conflict with user function + doc
disallow void argument types in user-defined functions
```

--------------------------------

### APIDOC: Multivariate Normal Vectorization and Mean Generalization

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces vectorization of outcome variates in multivariate normal distributions with efficiency boosts and generalizes the function to allow row vectors as means.

```APIDOC
vectorization of outcome variates in multivariate normal with efficiency boosts
generalization of multivariate normal to allow rwo vectors as means
```

--------------------------------

### APIDOC: User-Defined Function Definitions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces the ability to define user-defined functions directly within the Stan language, significantly enhancing model expressiveness and reusability.

```APIDOC
user-defined function definitions added to Stan language
```

--------------------------------

### APIDOC: std::cout Removal from API

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Initiates the removal of std::cout instances from the Stan API, aiming for cleaner output and better control over logging.

```APIDOC
removing many instances of std::cout from API (more still to go)
```

--------------------------------

### APIDOC: L-BFGS Optimization

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces L-BFGS as the new default optimization algorithm in Stan. This enhancement improves the efficiency and performance of model fitting.

```APIDOC
L-BFGS optimization (now the default)
```

--------------------------------

### Stan Language: Pochhammer Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Pochhammer functions, representing rising and falling factorials, were added to the Stan language for combinatorial and special function calculations.

```APIDOC
pochhammer_rising(x, n)
pochhammer_falling(x, n)
```

--------------------------------

### Stan Language: Rayleigh Distribution

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

The Rayleigh distribution was added to the Stan language, expanding the range of available probability distributions for modeling.

```APIDOC
rayleigh_lpdf(y | sigma)
rayleigh_lcdf(y | sigma)
rayleigh_lccdf(y | sigma)
rayleigh_rng(sigma)
```

--------------------------------

### APIDOC: multiply_lower_self_transpose Typo Fix

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Corrects a typo in the documentation for multiply_lower_self_transpose, ensuring accuracy in function references.

```APIDOC
fix typo in multiply_lower_self_transpose (thanks to Alexey Stukalov)
```

--------------------------------

### APIDOC: von Mises Random Number Generator

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds a new random number generator for the von Mises distribution, expanding the set of available statistical functions.

```APIDOC
added von Mises RNG
```

--------------------------------

### APIDOC: normal_cdf Range Enhancement

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Enhances the effective range of the normal_cdf function to prevent numerical underflow and overflow issues, improving its robustness.

```APIDOC
enhanced effective range of normal_cdf to prevent underflow/overflow
```

--------------------------------

### Stan Language: Sorting Indices Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

New `sort_indices_*()` functions were added to return indices in sorted order based on value, useful for custom sorting logic and data manipulation within Stan models.

```APIDOC
sort_indices_asc(array)
sort_indices_desc(array)
```

--------------------------------

### APIDOC: Reshaping Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Introduces new functions for reshaping data structures: to_matrix(), to_vector(), to_row_vector(), to_array_1d(), and to_array_2d(), facilitating data manipulation.

```APIDOC
reshaping functions: to_matrix(), to_vector(), to_row_vector(),
  to_array_1d(), to_array_2d()
```

--------------------------------

### Stan Language: Increment Log Probability Function

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

The `increment_log_prob()` function was introduced as the preferred method for incrementing the log probability, deprecating direct use of `lp__`. It also gained vectorized support for efficiency.

```APIDOC
increment_log_prob(value)
// Deprecates direct use of lp__
```

--------------------------------

### Stan Language: Bessel Functions

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Bessel functions of the first and second kind, both standard and modified, were added to the Stan language for advanced mathematical modeling.

```APIDOC
bessel_first_kind(nu, z)
bessel_second_kind(nu, z)
modified_bessel_first_kind(nu, z)
modified_bessel_second_kind(nu, z)
```

--------------------------------

### Stan Language: Multi-Normal Cholesky Random Number Generator

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Added `multi_normal_cholesky_rng()` for generating random numbers from a multi-normal distribution parameterized by its Cholesky factor, enhancing sampling capabilities.

```APIDOC
multi_normal_cholesky_rng(mu, L)
```

--------------------------------

### Stan Language: Cholesky Factor Data Type

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

A new `cholesky_factor` data type was introduced to the Stan language, allowing for more efficient and robust handling of Cholesky factors in models.

```APIDOC
cholesky_factor_cov[N]
```

--------------------------------

### New Stan Mathematical Functions (v1.3.0)

Source: https://github.com/stan-dev/stan/blob/develop/RELEASE-NOTES.txt

Adds several new mathematical functions to the Stan language, including specialized matrix operations and statistical functions like Owen's T function.

```APIDOC
Functions:
  mdivide_right_spd(matrix A, matrix B): Matrix division A * B^-1 for symmetric positive definite B
  mdivide_left_spd(matrix A, matrix B): Matrix division A^-1 * B for symmetric positive definite A
  owenst(real h, real a): Owen's T function
  columns_dot_product(matrix A, matrix B): Dot product of corresponding columns
  rows_dot_product(matrix A, matrix B): Dot product of corresponding rows
```
Tokens:
Raw