mlcolvar.core.stats.LDA

class mlcolvar.core.stats.LDA(in_features: int, n_states: int, mode: str = 'standard')

Bases: Stats

Fisher’s discriminant class.

Variables:

• in_features (int) – Number of input features

• out_features (int) – Number of ouput features (n_states - 1)

• n_states (int) – Number of states

• evals (torch.Tensor) – LDA eigenvalues

• evecs (torch.Tensor) – LDA eigenvectors

• S_b (torch.Tensor) – Between scatter matrix

• S_w (torch.Tensor) – Within scatter matrix

• sw_reg (float) – Regularization to S_w matrix

__init__(in_features: int, n_states: int, mode: str = 'standard')

Initialize a LDA object.

Parameters:

• in_features (int) – number of input features

• n_states (int) – number of states

• mode (string) – options: {‘standard’,’harmonic’}, standard or harmonic variant

Methods

__init__(in_features, n_states[, mode])	Initialize a LDA object.
compute(X, labels[, save_params])	Compute LDA eigenvalues and eigenvectors.
compute_scatter_matrices(X, labels[, ...])	Compute between scatter and within scatter matrices.
extra_repr()	Return the extra representation of the module.
forward(x)	Compute linear combination with saved eigenvectors

compute(X, labels, save_params=True)

Compute LDA eigenvalues and eigenvectors. First compute the scatter matrices S_between (S_b) and S_within (S_w) and then solve the generalized eigenvalue problem.

Parameters:

• X (array-like of shape (n_samples, in_features)) – Training data.

• labels (array-like of shape (n_samples,)) – states labels.

• save_params (bool, optional) – Whether to store parameters in model

Returns:

• eigvals (torch.Tensor) – LDA eigenvalues (n_states-1)

• eigvecs (torch.Tensor) – LDA eigenvectors (n_feature,n_states-1)

Notes

The eigenvecs object which is returned is a matrix whose column eigvecs[:,i] is the eigenvector associated to eigvals[i]

compute_scatter_matrices(X, labels, save_params=True)

Compute between scatter and within scatter matrices.

Parameters:

• X (array-like of shape (n_samples, in_features)) – Training data.

• labels (array-like of shape (n_samples,)) – states labels.

• save_params (bool, optional) – Whether to store parameters in model

Returns:

S_b,S_w – Between and within scatter matrices

Return type:

arrays of shape (in_features,in_features)

extra_repr() → str

Return the extra representation of the module.

To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable.

forward(x: Tensor) → Tensor

Compute linear combination with saved eigenvectors

Parameters:

x (torch.Tensor) – input

Returns:

out – output

Return type:

torch.Tensor

Attributes

T_destination
call_super_init
dump_patches
training