evox.algorithms.so.pso_variants.fs_pso

Module Contents

Classes

FSPSO

The Feature Selection PSO algorithm.

API

class evox.algorithms.so.pso_variants.fs_pso.FSPSO(pop_size: int, lb: torch.Tensor, ub: torch.Tensor, inertia_weight: float = 0.6, cognitive_coefficient: float = 2.5, social_coefficient: float = 0.8, mean=None, stdev=None, mutate_rate: float = 0.01, device: torch.device | None = None)[source]

Bases: evox.core.Algorithm

The Feature Selection PSO algorithm.

Initialization

Initialize the FSPSO algorithm with the given parameters.

Parameters:

  • pop_size – The size of the population.

  • lb – The lower bounds of the particle positions. Must be a 1D tensor.

  • ub – The upper bounds of the particle positions. Must be a 1D tensor.

  • inertia_weight – The inertia weight. Defaults to 0.6.

  • cognitive_coefficient – The cognitive weight. Defaults to 2.5.

  • social_coefficient – The social weight. Defaults to 0.8.

  • mean – The mean of the normal distribution. Defaults to None.

  • stdev – The standard deviation of the normal distribution. Defaults to None.

  • mutate_rate – The mutation rate. Defaults to 0.01.

  • device – The device to use for the tensors. Defaults to None.

init_step()[source]
step()[source]

Perform a normal optimization step using FSPSO.