Struct RandomTimeWarpAugmenter

pub struct RandomTimeWarpAugmenter {
    pub name: String,
    pub window_size: usize,
    pub speed_ratio_range: (f64, f64),
    /* private fields */
}

Augmenter that applies random time warping to the dataset This augmenter randomly selects a window of the time series, specified by the window_size argument and applies a speed change to it. The speed change is defined by the speed_ratio_range argument, which specifies the minimum and maximum speed ratio. The speed ratio is a multiplier that affects how fast or slow the selected window is stretched or compressed. If the window size is 0 or larger than the time series length, the entire series is warped.

Fields

name: String

window_size: usize

Length of the window to warp - a window of this size will be selected randomly for every time series in the dataset

speed_ratio_range: (f64, f64)

Range for random speed ratio: [min, max]

Implementations

impl RandomTimeWarpAugmenter

pub fn new(window_size: usize, speed_ratio_range: (f64, f64)) -> Self

Create a new augmenter with given window size. speed_ratio_range defines the min and max speed change (e.g. (0.5, 2.0)).

Trait Implementations

impl Augmenter for RandomTimeWarpAugmenter

fn augment_one(&self, x: &[f64]) -> Vec<f64>

Augment one time series

fn get_probability(&self) -> f64

Get the probability that this augmenter will augment a series in a batch

fn set_probability(&mut self, probability: f64)

By setting a probability with this function the augmenter will only augment a series in a batch with the specified probability

fn get_name(&self) -> String

fn augment_batch(&self, input: &mut Dataset, parallel: bool, per_sample: bool)

where Self: Sync,

Augment a whole batch

fn supports_per_sample(&self) -> bool

Indicate whether this augmenter supports per-sample chaining. By default, return true. Augmenters that need a batch level view should override this to return false.