Struct changepoint::Bocpd

pub struct Bocpd<X, Fx, Pr>where
    Fx: Rv<X> + HasSuffStat<X>,
    Pr: ConjugatePrior<X, Fx>,
    Fx::Stat: Clone,{ /* private fields */ }

Online Bayesian Change Point Detection state container.

Implementations

impl<X, Fx, Pr> Bocpd<X, Fx, Pr>where Fx: Rv<X> + HasSuffStat<X>, Pr: ConjugatePrior<X, Fx, Posterior = Pr> + Clone, Fx::Stat: Clone,

pub fn new(hazard_lambda: f64, predictive_prior: Pr) -> Self

Create a new Bocpd analyzer

Parameters

• hazard - 1/ hazard is the probability of the next step being a change-point. Therefore, the larger the value, the lower the prior probability is for the any point to be a change-point. Mean run-length is lambda - 1.
• predictive_prior - Prior for the predictive distribution. This should match your observed data.

Example

use changepoint::Bocpd;
use rv::prelude::*;

let cpd = Bocpd::new(
    250.0,
    NormalGamma::new_unchecked(0.0, 1.0, 1.0, 1.0),
);

pub fn reset_with_prior(&mut self, predictive_prior: Pr)

Reset the introspector and replace the predictive prior

pub fn collapse_stats(self) -> Self

Reduce the observed values into a new BOCPD with those observed values integrated into the prior.

Trait Implementations

impl<X, Fx, Pr> BocpdLike<X> for Bocpd<X, Fx, Pr>where Fx: Rv<X> + HasSuffStat<X>, Pr: ConjugatePrior<X, Fx, Posterior = Pr> + Clone, Fx::Stat: Clone,

fn step(&mut self, data: &X) -> &[f64]

Update the model with a new datum and return the distribution of run lengths.

type Fx = Fx

Type of type of distribution

type PosteriorPredictive = Mixture<Pr>

Type of predictive prior distribution

fn reset(&mut self)

Reset internal state, new run-lengths will refer to steps after this point.

fn pp(&self) -> Self::PosteriorPredictive

Generate the posterior predictive distribution

fn preload(&mut self, data: &[X])

Preload a seqeunce into the default suff stat

impl<X: Clone, Fx, Pr> Clone for Bocpd<X, Fx, Pr>where Fx: Rv<X> + HasSuffStat<X> + Clone, Pr: ConjugatePrior<X, Fx> + Clone, Fx::Stat: Clone + Clone,

fn clone(&self) -> Bocpd<X, Fx, Pr>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<X: Debug, Fx, Pr> Debug for Bocpd<X, Fx, Pr>where Fx: Rv<X> + HasSuffStat<X> + Debug, Pr: ConjugatePrior<X, Fx> + Debug, Fx::Stat: Clone + Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<X: PartialEq, Fx, Pr> PartialEq<Bocpd<X, Fx, Pr>> for Bocpd<X, Fx, Pr>where Fx: Rv<X> + HasSuffStat<X> + PartialEq, Pr: ConjugatePrior<X, Fx> + PartialEq, Fx::Stat: Clone + PartialEq,

fn eq(&self, other: &Bocpd<X, Fx, Pr>) -> bool

This method tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

This method tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl<X, Fx, Pr> StructuralPartialEq for Bocpd<X, Fx, Pr>where Fx: Rv<X> + HasSuffStat<X>, Pr: ConjugatePrior<X, Fx>, Fx::Stat: Clone,