Struct light_curve_feature::Periodogram

pub struct Periodogram<T: Float> { /* fields omitted */ }

A number of features based on Lomb–Scargle periodogram

Periodogram $P(\omega)$ is an estimate of spectral density of unevenly time series. Periodogram::new's peaks argument corresponds to a number of the most significant spectral density peaks to return. For each peak its period and "signal to noise" ratio is returned.

$$ \mathrm{signal~to~noise~of~peak} \equiv \frac{P(\omega_\mathrm{peak}) - \langle P(\omega) \rangle}{\sigma_{P(\omega)}}. $$

Periodogram can accept another dyn FeatureEvaluator for feature extraction from periodogram as it was time series without observation errors. You can even pass one Periodogram to another one if you are crazy enough

• Depends on: time, magnitude
• Minimum number of observations: 2 (or as required by sub-features)
• Number of features: $2 \times \mathrm{peaks}~+...$

Implementations

impl<T> Periodogram<T> where
    T: Float, 

pub fn new(peaks: usize) -> Self

New Periodogram that finds given number of peaks

pub fn set_freq_resolution(&mut self, resolution: f32) -> &mut Self

Set frequency resolution

The larger frequency resolution allows to find peak period with better precision

pub fn set_max_freq_factor(&mut self, max_freq_factor: f32) -> &mut Self

Multiply maximum (Nyquist) frequency

Maximum frequency is Nyquist frequncy multiplied by this factor. The larger factor allows to find larger frequency and makes PeriodogramPowerFft more precise. However large frequencies can show false peaks

pub fn set_nyquist(&mut self, nyquist: Box<dyn NyquistFreq<T>>) -> &mut Self

Define Nyquist frequency

pub fn add_feature(
    &mut self,
    feature: Box<dyn FeatureEvaluator<T>>
) -> &mut Self

Extend a feature to extract from periodogram

pub fn set_periodogram_algorithm(
    &mut self,
    periodogram_power: fn() -> Box<dyn PeriodogramPower<T>>
) -> &mut Self

pub fn init_thread_local_fft_plan(n: &[usize])

pub fn power(&self, ts: &mut TimeSeries<'_, T>) -> Vec<T>

pub fn freq_power(&self, ts: &mut TimeSeries<'_, T>) -> (Vec<T>, Vec<T>)

Trait Implementations

impl<T: Clone + Float> Clone for Periodogram<T>

fn clone(&self) -> Periodogram<T>

Returns a copy of the value.

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

Performs copy-assignment from source.

impl<T: Debug + Float> Debug for Periodogram<T>

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

Formats the value using the given formatter.

impl<T> Default for Periodogram<T> where
    T: Float, 

fn default() -> Self

Returns the "default value" for a type.

impl<T> FeatureEvaluator<T> for Periodogram<T> where
    T: Float, 

fn eval(&self, ts: &mut TimeSeries<'_, T>) -> Result<Vec<T>, EvaluatorError>

Should return the vector of feature values

fn eval_or_fill(&self, ts: &mut TimeSeries<'_, T>, fill_value: T) -> Vec<T>

fn get_info(&self) -> &EvaluatorInfo

Get feature evaluator meta-information

fn get_names(&self) -> Vec<&str>

Should return the vector of feature names. The length and feature order should correspond to eval() output

fn size_hint(&self) -> usize

Should return the size of vectors returned by eval() and get_names()

fn min_ts_length(&self) -> usize

Should return minimum time series length to successfully find feature value

fn is_t_required(&self) -> bool

fn is_m_required(&self) -> bool

fn is_w_required(&self) -> bool

fn is_sorting_required(&self) -> bool

fn check_ts_length(
    &self,
    ts: &TimeSeries<'_, T>
) -> Result<usize, EvaluatorError>