Struct EvocEmbeddingParams 

pub struct EvocEmbeddingParams<T> {
    pub n_label_prop_iter: usize,
    pub base_init_threshold: usize,
    pub scaling: T,
    pub n_epochs: usize,
    pub noise_level: T,
    pub negative_sample_rate: T,
    pub initial_alpha: T,
    pub beta1: T,
    pub beta2: T,
    pub eps: T,
}

Parameters for the EVoC node embedding

Fields

n_label_prop_iter: usize

Number of iterations during the label propagation

base_init_threshold: usize

Base initialisation threshold

scaling: T

Scaling parameter

n_epochs: usize

Number of training epochs. Higher values give more refined embeddings at the cost of runtime. Default: 50.

noise_level: T

Scaling factor in the attractive gradient kernel numerator ( -2 * noise_level * d - 2). Controls how aggressively connected nodes attract. Default: 0.5.

negative_sample_rate: T

Controls how many negative (repulsive) samples are drawn per positive edge sample. Lower values reduce repulsion; setting to zero effectively disables it. Default: 1.0.

initial_alpha: T

Initial learning rate, linearly decayed to zero over training. Default: 0.1.

beta1: T

Adam first moment decay rate. Default: 0.5.

beta2: T

Adam second moment decay rate. Default: 0.9.

eps: T

Adam numerical stability constant. Default: 1e-7.

Trait Implementations

impl<T: Clone> Clone for EvocEmbeddingParams<T>

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

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<T: Debug> Debug for EvocEmbeddingParams<T>

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

Formats the value using the given formatter.

impl<T: EvocFloat> Default for EvocEmbeddingParams<T>

Defaults for the EvocEmbeddingParameters

fn default() -> Self

Returns the “default value” for a type.