pulsatrix 1.1.0
Loading...
Searching...
No Matches
pulsatrix::Anchor Struct Reference

FindAnchor()'s result. More...

#include <anchors.hpp>

Public Member Functions

bool applies_to (const std::vector< float > &x) const
 Whether x (flattened) satisfies every predicate.
 

Public Attributes

int64_t predicted_class = 0
 The class the model gives the input (its largest output), which the anchor keeps.
 
std::vector< AnchorPredicate > predicates
 The rule, one interval per feature, in the order the search first constrained each. Empty: the model gives this class almost everywhere.
 
double precision = 0.0
 The share of sampled inputs satisfying the rule that kept the class.
 
double precision_lower = 0.0
 Its KL confidence bounds at 1 - delta.
 
double precision_upper = 1.0
 
int64_t samples = 0
 Samples behind the estimate.
 
double coverage = 0.0
 The share of the background the rule applies to.
 
bool meets_threshold = false
 Whether it reached the threshold. When no anchor does, this is the most precise candidate found, and false.
 
int64_t model_calls = 0
 Calls to the model over the whole search.
 

Detailed Description

FindAnchor()'s result.

Member Function Documentation

◆ applies_to()

bool pulsatrix::Anchor::applies_to ( const std::vector< float > &  x) const

Whether x (flattened) satisfies every predicate.

Member Data Documentation

◆ coverage

double pulsatrix::Anchor::coverage = 0.0

The share of the background the rule applies to.

◆ meets_threshold

bool pulsatrix::Anchor::meets_threshold = false

Whether it reached the threshold. When no anchor does, this is the most precise candidate found, and false.

◆ model_calls

int64_t pulsatrix::Anchor::model_calls = 0

Calls to the model over the whole search.

◆ precision

double pulsatrix::Anchor::precision = 0.0

The share of sampled inputs satisfying the rule that kept the class.

◆ precision_lower

double pulsatrix::Anchor::precision_lower = 0.0

Its KL confidence bounds at 1 - delta.

◆ precision_upper

double pulsatrix::Anchor::precision_upper = 1.0

◆ predicates

std::vector<AnchorPredicate> pulsatrix::Anchor::predicates

The rule, one interval per feature, in the order the search first constrained each. Empty: the model gives this class almost everywhere.

◆ predicted_class

int64_t pulsatrix::Anchor::predicted_class = 0

The class the model gives the input (its largest output), which the anchor keeps.

◆ samples

int64_t pulsatrix::Anchor::samples = 0

Samples behind the estimate.


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