fastlowess-cpp
High-performance LOWESS smoothing
Loading...
Searching...
No Matches
fastlowess::OnlineOutput Class Reference

Result of a single online update step. More...

#include <fastlowess.hpp>

Public Member Functions

bool has_value () const
 True when the window has enough points to produce a smoothed estimate.
 
bool dropped () const
 True when missing = "drop" ignored a non-finite point.
 
OnlinePointStatus status () const
 Detailed outcome of the latest add_point call.
 
double y () const
 Smoothed value for the latest point (valid only when has_value() == true).
 
double standard_error () const
 Standard error (NaN if not computed).
 
double residual () const
 Residual y − smoothed (NaN if not computed).
 
double robustness_weight () const
 Robustness weight for the latest point (NaN if not computed).
 
int iterations_used () const
 Number of robustness iterations performed (−1 if not applicable).
 
double derivative () const
 Local fit derivative (slope) for the latest point (NaN if not computed).
 
double confidence_lower () const
 Lower confidence interval bound for the latest point (NaN if not computed).
 
double confidence_upper () const
 Upper confidence interval bound for the latest point (NaN if not computed).
 
double prediction_lower () const
 Lower prediction interval bound for the latest point (NaN if not computed).
 
double prediction_upper () const
 Upper prediction interval bound for the latest point (NaN if not computed).
 

Friends

class OnlineLowess
 
template<typename U >
class Expected
 

Detailed Description

Result of a single online update step.

Call status() / has_value() / dropped() to distinguish a ready estimate, a still-filling window, and a point dropped under missing = "drop". All optional fields are NaN when not computed.

Member Function Documentation

◆ confidence_lower()

double fastlowess::OnlineOutput::confidence_lower ( ) const
inline

Lower confidence interval bound for the latest point (NaN if not computed).

◆ confidence_upper()

double fastlowess::OnlineOutput::confidence_upper ( ) const
inline

Upper confidence interval bound for the latest point (NaN if not computed).

◆ derivative()

double fastlowess::OnlineOutput::derivative ( ) const
inline

Local fit derivative (slope) for the latest point (NaN if not computed).

◆ dropped()

bool fastlowess::OnlineOutput::dropped ( ) const
inline

True when missing = "drop" ignored a non-finite point.

◆ has_value()

bool fastlowess::OnlineOutput::has_value ( ) const
inline

True when the window has enough points to produce a smoothed estimate.

◆ iterations_used()

int fastlowess::OnlineOutput::iterations_used ( ) const
inline

Number of robustness iterations performed (−1 if not applicable).

◆ prediction_lower()

double fastlowess::OnlineOutput::prediction_lower ( ) const
inline

Lower prediction interval bound for the latest point (NaN if not computed).

◆ prediction_upper()

double fastlowess::OnlineOutput::prediction_upper ( ) const
inline

Upper prediction interval bound for the latest point (NaN if not computed).

◆ residual()

double fastlowess::OnlineOutput::residual ( ) const
inline

Residual y − smoothed (NaN if not computed).

◆ robustness_weight()

double fastlowess::OnlineOutput::robustness_weight ( ) const
inline

Robustness weight for the latest point (NaN if not computed).

◆ standard_error()

double fastlowess::OnlineOutput::standard_error ( ) const
inline

Standard error (NaN if not computed).

◆ status()

OnlinePointStatus fastlowess::OnlineOutput::status ( ) const
inline

Detailed outcome of the latest add_point call.

◆ y()

double fastlowess::OnlineOutput::y ( ) const
inline

Smoothed value for the latest point (valid only when has_value() == true).

Friends And Related Symbol Documentation

◆ Expected

template<typename U >
friend class Expected
friend

◆ OnlineLowess


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