std::numeric_limits< boost::probabilities::likelihood< Domain, Value > > Struct Template Reference

Specialization of std::numeric_limits for likelihoods. More...

#include <likelihoods.hpp>

List of all members.

Static Public Member Functions
static Value	min () throw ()
static Value	max () throw ()
static Value	epsilon () throw ()
static Value	round_error () throw ()
static Value	infinity () throw ()
static Value	quiet_NaN () throw ()
static Value	signaling_NaN () throw ()
static Value	denorm_min () throw ()
Static Public Attributes
static const bool	is_specialized = true
static const int	digits = numeric_limits<Value>::digits
static const int	digits10 = numeric_limits<Value>::digits10
static const bool	is_signed = false
static const bool	is_integer = numeric_limits<Value>::is_integer
static const bool	is_exact = numeric_limits<Value>::is_exact
static const int	radix = numeric_limits<Value>::radix
static const int	min_exponent = numeric_limits<Value>::min_exponent
static const int	min_exponent10 = numeric_limits<Value>::min_exponent10
static const int	max_exponent = numeric_limits<Value>::max_exponent
static const int	max_exponent10 = numeric_limits<Value>::max_exponent10
static const bool	has_infinity = numeric_limits<Value>::has_infinity
static const bool	has_quiet_NaN = numeric_limits<Value>::has_quiet_NaN
static const bool	has_signaling_NaN = numeric_limits<Value>::has_signaling_NaN
static const float_denorm_style	has_denorm = numeric_limits<Value>::has_denorm
static const bool	has_denorm_loss = numeric_limits<Value>::has_denorm_loss
static const bool	is_iec559 = numeric_limits<Value>::is_iec559
static const bool	is_bounded = numeric_limits<Value>::is_bounded
static const bool	is_modulo = numeric_limits<Value>::is_modulo
static const bool	traps = numeric_limits<Value>::traps
static const bool	tinyness_before = numeric_limits<Value>::tinyness_before
static const float_round_style	round_style = numeric_limits<Value>::round_style

---

Detailed Description

template<typename Domain, typename Value>
struct std::numeric_limits< boost::probabilities::likelihood< Domain, Value > >

Specialization of std::numeric_limits for likelihoods.

In almost all instances, the functions and values encapsulated for std::numeric_limits<boost::probabilities::likelihood> are simply delegated to the corresponding functions and values for the underlying value_type. The following points are worthy of note.

• is_signed is false regardless of the value_type as likelihoods are considered to be proportional to probabilities with a positive constant of proportionality.

• min() returns the value of std::numeric_limits<Value>::min() for any domain. For types without denormalization, this corresponds to the minimum finite value representable by the internal type; for types with denormalization this corresponds to the minimum positive normalized value representable by the internal type. This seems to be the most appropriate interpretation of the intent of std:numeric_limits<>::min().

• max() returns the value of std::numeric_limits<Value>::max() for any domain. For types without denormalization, this corresponds to the maximum finite value representable by the internal type; for types with denormalization this corresponds to the maximum positive normalized value representable by the internal type. This seems to be the most appropriate interpretation of the intent of std:numeric_limits<>::max(). Note that this behavior is unlike the specialization for the boost::probabilities::probability class, which instead returns unity (1) as a consequence of the closed interval for probabilities.

---

The documentation for this struct was generated from the following file:

• likelihoods.hpp

---