waveletml.models package

waveletml.models.base_model module

class waveletml.models.base_model.BaseModel

Bases: BaseEstimator

Base class for all models

evaluate(y_true, y_pred, list_metrics=None)

Evaluate the model using specified metrics.

Parameters:

• y_true (array-like) – True target values.

• y_pred (array-like) – Model’s predicted values.

• list_metrics (list of str, optional) – Names of metrics for evaluation (e.g., “MSE”, “MAE”).

Returns:

Evaluation metrics and their values.

Return type:

dict

fit(X, y)

Fit the model to the training data.

static load_model(load_path='history', filename='model.pkl')

Load a model from a pickle file.

Parameters:

• load_path (str, optional) – Path to load the model from (default: “history”).

• filename (str, optional) – Filename of the saved model (default: “model.pkl”).

Returns:

The loaded model.

Return type:

BaseMlp

predict(X)

Predict using the trained model.

save_evaluation_metrics(y_true, y_pred, list_metrics=('RMSE', 'MAE'), save_path='history', filename='metrics.csv')

Save evaluation metrics to a CSV file.

Parameters:

• y_true (array-like) – Ground truth values.

• y_pred (array-like) – Model predictions.

• list_metrics (list of str, optional) – Metrics for evaluation (default: (“RMSE”, “MAE”)).

• save_path (str, optional) – Path to save the file (default: “history”).

• filename (str, optional) – Filename for saving metrics (default: “metrics.csv”).

save_model(save_path='history', filename='model.pkl')

Save the trained model to a pickle file.

Parameters:

• save_path (str, optional) – Path to save the model (default: “history”).

• filename (str, optional) – Filename for saving model, with “.pkl” extension (default: “model.pkl”).

save_training_loss(save_path='history', filename='loss.csv')

Save training loss history to a CSV file.

Parameters:

• save_path (str, optional) – Path to save the file (default: “history”).

• filename (str, optional) – Filename for saving loss history (default: “loss.csv”).

save_y_predicted(X, y_true, save_path='history', filename='y_predicted.csv')

Save true and predicted values to a CSV file.

Parameters:

• X (array-like or torch.Tensor) – Input features.

• y_true (array-like) – True values.

• save_path (str, optional) – Path to save the file (default: “history”).

• filename (str, optional) – Filename for saving predicted values (default: “y_predicted.csv”).

score(X, y)

Evaluate the model’s performance on the given data.

waveletml.models.custom_wnn module

class waveletml.models.custom_wnn.BaseCustomWNN(input_dim, hidden_dim, output_dim, wavelet_fn='morlet', act_output=None, seed=None)

Bases: Module

Base class for custom wavelet neural networks. This class is not meant to be used directly.

SUPPORTED_ACTIVATIONS = ['Threshold', 'ReLU', 'RReLU', 'Hardtanh', 'ReLU6', 'Sigmoid', 'Hardsigmoid', 'Tanh', 'SiLU', 'Mish', 'Hardswish', 'ELU', 'CELU', 'SELU', 'GLU', 'GELU', 'Hardshrink', 'LeakyReLU', 'LogSigmoid', 'Softplus', 'Softshrink', 'MultiheadAttention', 'PReLU', 'Softsign', 'Tanhshrink', 'Softmin', 'Softmax', 'Softmax2d', 'LogSoftmax']

SUPPORTED_WAVELETS = ['morlet', 'mexican_hat', 'haar', 'db1', 'db2', 'sym2', 'coif1', 'bior1.3', 'bior1.5', 'rbio1.3', 'rbio1.5', 'dmey', 'cmor']

forward(x)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_weights()

Retrieve network weights as a flattened array.

Returns:

Flattened array of the model’s weights.

Return type:

• np.ndarray

get_weights_size()

Calculate the total number of trainable parameters in the model.

Returns:

Total number of parameters.

Return type:

• int

set_weights(solution)

Set network weights based on a given solution vector.

Parameters:

solution (-) – A flat array of weights to set in the model.

training: bool

class waveletml.models.custom_wnn.CustomWaveletExpansionNetwork(input_dim, hidden_dim, output_dim, wavelet_fn='morlet', act_output=None, seed=None)

Bases: BaseCustomWNN

A custom wavelet neural network using a feature expansion-based wavelet layer.

This network computes the wavelet function for each input dimension, expands the feature space, and passes the result through a standard linear output layer.

In this version, we calculate the z value for each input dimension, then we apply the wavelet function to each z value. The output of each hidden neuron is forming a new feature space. Wavelet-based feature expansion. The output layer is a standard linear layer. (weights and bias).

The number of parameters for WNN(3, 5, 1) is: 5*3 (centers) + 5*3 (scales) + 3*5 (weights at output layer) + 1 (bias ouput) = 5*3 + 5*3 + 3*5 + 1 = 46

wavelet_layer

The wavelet-based feature expansion layer.

Type:

WaveletExpansionLayer

output_layer

The standard linear output layer.

Type:

torch.nn.Linear

forward(x)

Performs the forward pass of the network.

forward(x)

Performs the forward pass of the network.

Parameters:

x (torch.Tensor) – Input tensor of shape (batch_size, input_dim).

Returns:

Output tensor of shape (batch_size, output_dim).

Return type:

torch.Tensor

training: bool

class waveletml.models.custom_wnn.CustomWaveletProductNetwork(input_dim, hidden_dim, output_dim, wavelet_fn='morlet', act_output=None, seed=None)

Bases: BaseCustomWNN

A custom wavelet neural network using a product-based wavelet layer. This network computes the wavelet function for each input dimension, takes the product of the results for each hidden neuron, and passes the result through a standard linear output layer.

In this version, we calculate the z value for each input dimension, then we apply the wavelet function to each z value. The output of each hidden neuron is the product of all wavelet functions from input to that hidden neuron. The output layer is a standard linear layer. (weights and bias).

Examples: The number of parameters for WNN(3, 5, 1) is: 5*3 (centers) + 5*3 (scales) + 5 (weights at output layer) + 1 (bias ouput) = 5*3 + 5*3 + 5 + 1 = 36

wavelet_layer

The wavelet-based product layer.

Type:

WaveletProductLayer

output_layer

The standard linear output layer.

Type:

torch.nn.Linear

forward(x)

Performs the forward pass of the network.

forward(x)

Performs the forward pass of the network.

Parameters:

x (torch.Tensor) – Input tensor of shape (batch_size, input_dim).

Returns:

Output tensor of shape (batch_size, output_dim).

Return type:

torch.Tensor

training: bool

class waveletml.models.custom_wnn.CustomWaveletSummationNetwork(input_dim, hidden_dim, output_dim, wavelet_fn='morlet', act_output=None, seed=None)

Bases: BaseCustomWNN

A custom wavelet neural network using a summation-based wavelet layer.

This network computes the wavelet function for each input dimension, takes the sum of the results for each hidden neuron, and passes the result through a standard linear output layer.

In this version, we calculate the z value for each input dimension, then we apply the wavelet function to each z value. The output of each hidden neuron is the sum of all wavelet functions from input to that hidden neuron. The output layer is a standard linear layer. (weights and bias).

The number of parameters for WNN(3, 5, 1) is: 5*3 (centers) + 5*3 (scales) + 5 (weights at output layer) + 1 (bias ouput) = 5*3 + 5*3 + 5 + 1 = 36

wavelet_layer

The wavelet-based summation layer.

Type:

WaveletSummationLayer

output_layer

The standard linear output layer.

Type:

torch.nn.Linear

forward(x)

Performs the forward pass of the network.

forward(x)

Performs the forward pass of the network.

Parameters:

x (torch.Tensor) – Input tensor of shape (batch_size, input_dim).

Returns:

Output tensor of shape (batch_size, output_dim).

Return type:

torch.Tensor

training: bool

class waveletml.models.custom_wnn.CustomWaveletWeightedLinearNetwork(input_dim, hidden_dim, output_dim, wavelet_fn='morlet', act_output=None, seed=None)

Bases: BaseCustomWNN

A custom wavelet neural network using a weighted linear layer.

In this version, we calculate the sum of all inputs to each neuron (wx) Then we calculate z = (wx - b) / a, where wx is the weighted sum of inputs. Then we apply the wavelet function to the z value. The output layer is a standard linear layer. (weights and bias)

Examples: The number of parameters for WNN(3, 5, 1) is: 3 * 5 (weights) + 5 (centers) + 5 (scales) + 5 (weights at output layer) + 1 (bias ouput) = 3 * 5 + 5 + 5 + 1 = 26

wavelet_layer

The wavelet-based weighted linear layer.

Type:

WaveletWeightedLinearLayer

output_layer

The standard linear output layer.

Type:

torch.nn.Linear

forward(x)

Performs the forward pass of the network.

forward(x)

Performs the forward pass of the network.

Parameters:

x (torch.Tensor) – Input tensor of shape (batch_size, input_dim).

Returns:

Output tensor of shape (batch_size, output_dim).

Return type:

torch.Tensor

training: bool

waveletml.models.gd_wnn module

class waveletml.models.gd_wnn.BaseGdWnnModel(size_hidden=10, wavelet_fn='morlet', act_output=None, epochs=1000, batch_size=16, optim='Adam', optim_params=None, valid_rate=0.1, seed=42, verbose=True, device=None, **kwargs)

Bases: BaseModel

Base class for Fully Gradient-based Wavelet Neural Network (GdWNN) models.

This class provides common functionality for both classifiers and regressors, including data processing, training, and callback handling.

size_hidden

Number of hidden neurons in the wavelet neural network.

Type:

int

wavelet_fn

Name of the wavelet function to use.

Type:

str

act_output

Activation function for the output layer.

Type:

callable or None

epochs

Number of training epochs.

Type:

int

batch_size

Batch size for training.

Type:

int

optim

Name of the optimizer to use.

Type:

str

optim_params

Parameters for the optimizer.

Type:

dict

valid_rate

Proportion of data to use for validation.

Type:

float

seed

Random seed for reproducibility.

Type:

int

verbose

Whether to print training progress.

Type:

bool

device

Device to use for training (‘cpu’ or ‘gpu’).

Type:

str

callbacks

List of callback instances for training.

Type:

list

wnn_model

Wavelet neural network model class.

Type:

class

kwargs

Additional keyword arguments.

Type:

dict

size_input

Number of input features.

Type:

int or None

size_output

Number of output features.

Type:

int or None

network

Neural network instance.

Type:

torch.nn.Module or None

optimizer

Optimizer instance.

Type:

torch.optim.Optimizer or None

criterion

Loss function.

Type:

torch.nn.Module or None

valid_mode

Whether validation mode is enabled.

Type:

bool

loss_train

List of training losses for each epoch.

Type:

list

_process_data(X, y)

Processes the input data for training.

_train(X, y)

Trains the model on the given data.

class waveletml.models.gd_wnn.GdWnnClassifier(size_hidden=10, wavelet_fn='morlet', act_output=None, epochs=1000, batch_size=16, optim='Adam', optim_params=None, valid_rate=0.1, seed=42, verbose=True, device=None, **kwargs)

Bases: BaseGdWnnModel, ClassifierMixin

Gradient-based Wavelet Neural Network (GdWNN) Classifier.

This class implements a wavelet-based neural network for classification tasks, leveraging gradient-based optimization for training.

Parameters:

size_hiddenint, optional

Number of hidden neurons in the wavelet neural network (default is 10).

wavelet_fnstr, optional

Name of the wavelet function to use (default is “morlet”).

act_outputcallable or None, optional

Activation function for the output layer (default is None).

epochsint, optional

Number of training epochs (default is 1000).

batch_sizeint, optional

Batch size for training (default is 16).

optimstr, optional

Name of the optimizer to use (default is “Adam”).

optim_paramsdict, optional

Parameters for the optimizer (default is None).

valid_ratefloat, optional

Proportion of data to use for validation (default is 0.1).

seedint, optional

Random seed for reproducibility (default is 42).

verbosebool, optional

Whether to print training progress (default is True).

devicestr, optional

Device to use for training (‘cpu’ or ‘gpu’, default is None).

kwargsdict, optional

Additional keyword arguments.

size_hidden

Number of hidden neurons in the wavelet neural network.

Type:

int

wavelet_fn

Name of the wavelet function to use.

Type:

str

act_output

Activation function for the output layer.

Type:

callable or None

epochs

Number of training epochs.

Type:

int

batch_size

Batch size for training.

Type:

int

optim

Name of the optimizer to use.

Type:

str

optim_params

Parameters for the optimizer.

Type:

dict

valid_rate

Proportion of data to use for validation.

Type:

float

seed

Random seed for reproducibility.

Type:

int

verbose

Whether to print training progress.

Type:

bool

device

Device to use for training (‘cpu’ or ‘gpu’).

Type:

str

classes_

Unique class labels in the dataset.

Type:

np.ndarray or None

task

Classification task type (‘classification’ or ‘binary_classification’).

Type:

str

_process_data(X, y)

Processes the input data for training and validation.

fit(X, y)

Fits the model to the training data.

predict(X)

Predicts class labels for the input data.

predict_proba(X)

Predicts class probabilities for the input data.

score(X, y)

Computes the accuracy of the model on the given data.

evaluate(y_true, y_pred, list_metrics)

Evaluates the model using specified metrics.

scores(X, y, list_metrics)

Computes evaluation metrics on the given data.

evaluate(y_true, y_pred, list_metrics=('AS', 'RS'))

Evaluates the model using specified metrics.

fit(X, y)

Fits the GdWnnClassifier to the training data.

predict(X)

Predict class labels for the input data.

predict_proba(X)

Predict class probabilities for the input data.

score(X, y)

Return the accuracy on the given test data and labels.

scores(X, y, list_metrics=('AS', 'RS'))

Computes evaluation metrics on the given data.

class waveletml.models.gd_wnn.GdWnnRegressor(size_hidden=10, wavelet_fn='morlet', act_output=None, epochs=1000, batch_size=16, optim='Adam', optim_params=None, valid_rate=0.1, seed=42, verbose=True, device=None, **kwargs)

Bases: BaseGdWnnModel, RegressorMixin

Gradient-based Wavelet Neural Network (GdWNN) Regressor.

This class implements a wavelet-based neural network for regression tasks, leveraging gradient-based optimization for training.

Parameters:

• size_hidden (int, optional) – Number of hidden neurons in the wavelet neural network (default is 10).

• wavelet_fn (str, optional) – Name of the wavelet function to use (default is “morlet”).

• act_output (callable or None, optional) – Activation function for the output layer (default is None).

• epochs (int, optional) – Number of training epochs (default is 1000).

• batch_size (int, optional) – Batch size for training (default is 16).

• optim (str, optional) – Name of the optimizer to use (default is “Adam”).

• optim_params (dict, optional) – Parameters for the optimizer (default is None).

• valid_rate (float, optional) – Proportion of data to use for validation (default is 0.1).

• seed (int, optional) – Random seed for reproducibility (default is 42).

• verbose (bool, optional) – Whether to print training progress (default is True).

• device (str, optional) – Device to use for training (‘cpu’ or ‘gpu’, default is None).

• kwargs (dict, optional) – Additional keyword arguments.

size_hidden

Number of hidden neurons in the wavelet neural network.

Type:

int

wavelet_fn

Name of the wavelet function to use.

Type:

str

act_output

Activation function for the output layer.

Type:

callable or None

epochs

Number of training epochs.

Type:

int

batch_size

Batch size for training.

Type:

int

optim

Name of the optimizer to use.

Type:

str

optim_params

Parameters for the optimizer.

Type:

dict

valid_rate

Proportion of data to use for validation.

Type:

float

seed

Random seed for reproducibility.

Type:

int

verbose

Whether to print training progress.

Type:

bool

device

Device to use for training (‘cpu’ or ‘gpu’).

Type:

str

task

Regression task type (‘regression’ or ‘multi_regression’).

Type:

str

size_input

Number of input features.

Type:

int or None

size_output

Number of output features.

Type:

int or None

network

Neural network instance.

Type:

torch.nn.Module or None

optimizer

Optimizer instance.

Type:

torch.optim.Optimizer or None

criterion

Loss function.

Type:

torch.nn.Module or None

valid_mode

Whether validation mode is enabled.

Type:

bool

loss_train

List of training losses for each epoch.

Type:

list

_process_data(X, y)

Processes the input data for training and validation.

fit(X, y)

Fits the model to the training data.

predict(X)

Predicts the output values for the provided input data.

score(X, y)

Computes the R2 score of the model on the given data.

evaluate(y_true, y_pred, list_metrics=('MSE', 'MAE'))

Evaluates the model using specified metrics.

scores(X, y, list_metrics=('MSE', 'MAE'))

Computes evaluation metrics on the given data.

evaluate(y_true, y_pred, list_metrics=('MSE', 'MAE'))

Evaluates the model using specified metrics for regression tasks.

fit(X, y)

Fits the GdWnnRegressor to the training data.

predict(X)

Predicts the output for the given input data.

score(X, y)

Returns the coefficient of determination R2 of the prediction.

scores(X, y, list_metrics=('MSE', 'MAE'))

Computes evaluation metrics on the given data.

waveletml.models.mha_wnn module

class waveletml.models.mha_wnn.BaseMhaWnnModel(size_hidden=10, wavelet_fn='morlet', act_output=None, optim='Adam', optim_params=None, obj_name=None, seed=42, verbose=True, wnn_type=None, lb=None, ub=None, mode='single', n_workers=None, termination=None)

Bases: BaseModel

Base class for Metaheuristic-Optimized Wavelet Neural Network (MhaWNN) models.

This class serves as a foundation for constructing Wavelet Neural Networks (WNNs) optimized using various metaheuristic algorithms from the Mealpy library. It supports both regression and classification tasks, offering flexible model configuration, optimization setup, and training management.

Parameters:

• size_hidden (int, optional (default=10)) – Number of hidden neurons in the WNN.

• wavelet_fn (str, optional (default="morlet")) – Name of the wavelet basis function used in hidden layers.

• act_output (callable or None, optional (default=None)) – Activation function for the output layer.

• optim (str, optional (default="Adam")) – Name of the metaheuristic optimizer. Must be supported by Mealpy.

• optim_params (dict, optional (default=None)) – Dictionary of parameters to configure the optimizer.

• obj_name (str or None, optional (default=None)) – Name of the objective function or performance metric (e.g., “mse”, “accuracy”).

• seed (int, optional (default=42)) – Random seed for reproducibility.

• verbose (bool, optional (default=True)) – If True, prints optimization progress during training.

• wnn_type (str or subclass of BaseCustomWNN, optional (default=None)) – Type or custom implementation of the Wavelet Neural Network architecture.

• lb (float, int, list, or np.ndarray, optional (default=None)) – Lower bounds for the model’s trainable parameters.

• ub (float, int, list, or np.ndarray, optional (default=None)) – Upper bounds for the model’s trainable parameters.

• mode (str, optional (default='single')) – Mode for optimization (‘single’ or ‘swarm’, ‘thread’ or ‘process’).

• n_workers (int or None, optional (default=None)) – Number of parallel workers used in optimizer.

• termination (dict or callable, optional (default=None)) – Termination condition for the optimization process.

size_input

Number of features in the input dataset.

Type:

int or None

size_output

Number of output targets.

Type:

int or None

network

Constructed wavelet neural network.

Type:

torch.nn.Module or None

optimizer

Metaheuristic optimizer instance.

Type:

Optimizer or None

metric_class

Metric computation class based on selected objective.

Type:

callable or None

loss_train

History of training loss over optimization iterations.

Type:

list

minmax

Indicates whether the optimization is minimizing or maximizing the objective.

Type:

str or None

data

Training data (X, y) used during optimization.

Type:

tuple

_set_optimizer(optim=None, optim_params=None)

Initializes the optimizer using a string or an Optimizer instance.

get_name()

Returns a string describing the model and optimizer configuration.

build_model()

Constructs the WNN architecture, initializes optimizer and loss functions.

_set_lb_ub(lb=None, ub=None, n_dims=None)

Normalizes and validates lower and upper bounds for optimization.

objective_function(solution=None)

Evaluates the loss/fitness of the model given a parameter solution.

_fit(X, y)

Executes the training process using metaheuristic optimization on (X, y).

SUPPORTED_CLS_OBJECTIVES = {'AS': 'max', 'BSL': 'min', 'CEL': 'min', 'CKS': 'max', 'F1S': 'max', 'F2S': 'max', 'FBS': 'max', 'GINI': 'min', 'GMS': 'max', 'HL': 'min', 'HS': 'max', 'JSI': 'max', 'KLDL': 'min', 'LS': 'max', 'MCC': 'max', 'NPV': 'max', 'PS': 'max', 'ROC-AUC': 'max', 'RS': 'max', 'SS': 'max'}

SUPPORTED_OPTIMIZERS = ['OriginalABC', 'OriginalACOR', 'AugmentedAEO', 'EnhancedAEO', 'ImprovedAEO', 'ModifiedAEO', 'OriginalAEO', 'MGTO', 'OriginalAGTO', 'DevALO', 'OriginalALO', 'OriginalAO', 'OriginalAOA', 'IARO', 'LARO', 'OriginalARO', 'OriginalASO', 'OriginalAVOA', 'OriginalArchOA', 'AdaptiveBA', 'DevBA', 'OriginalBA', 'DevBBO', 'OriginalBBO', 'OriginalBBOA', 'OriginalBES', 'ABFO', 'OriginalBFO', 'OriginalBMO', 'DevBRO', 'OriginalBRO', 'OriginalBSA', 'ImprovedBSO', 'OriginalBSO', 'CleverBookBeesA', 'OriginalBeesA', 'ProbBeesA', 'OriginalCA', 'OriginalCDO', 'OriginalCEM', 'OriginalCGO', 'DevCHIO', 'OriginalCHIO', 'OriginalCOA', 'OCRO', 'OriginalCRO', 'OriginalCSA', 'OriginalCSO', 'OriginalCircleSA', 'OriginalCoatiOA', 'JADE', 'OriginalDE', 'SADE', 'SAP_DE', 'DevDMOA', 'OriginalDMOA', 'OriginalDO', 'DevEFO', 'OriginalEFO', 'OriginalEHO', 'AdaptiveEO', 'ModifiedEO', 'OriginalEO', 'OriginalEOA', 'LevyEP', 'OriginalEP', 'CMA_ES', 'LevyES', 'OriginalES', 'Simple_CMA_ES', 'OriginalESOA', 'OriginalEVO', 'OriginalFA', 'DevFBIO', 'OriginalFBIO', 'OriginalFFA', 'OriginalFFO', 'OriginalFLA', 'DevFOA', 'OriginalFOA', 'WhaleFOA', 'DevFOX', 'OriginalFOX', 'OriginalFPA', 'BaseGA', 'EliteMultiGA', 'EliteSingleGA', 'MultiGA', 'SingleGA', 'OriginalGBO', 'DevGCO', 'OriginalGCO', 'OriginalGJO', 'OriginalGOA', 'DevGSKA', 'OriginalGSKA', 'Matlab101GTO', 'Matlab102GTO', 'OriginalGTO', 'GWO_WOA', 'IGWO', 'OriginalGWO', 'RW_GWO', 'OriginalHBA', 'OriginalHBO', 'OriginalHC', 'SwarmHC', 'OriginalHCO', 'OriginalHGS', 'OriginalHGSO', 'OriginalHHO', 'DevHS', 'OriginalHS', 'OriginalICA', 'OriginalINFO', 'OriginalIWO', 'DevJA', 'LevyJA', 'OriginalJA', 'DevLCO', 'ImprovedLCO', 'OriginalLCO', 'OriginalMA', 'OriginalMFO', 'OriginalMGO', 'OriginalMPA', 'OriginalMRFO', 'WMQIMRFO', 'OriginalMSA', 'DevMVO', 'OriginalMVO', 'OriginalNGO', 'ImprovedNMRA', 'OriginalNMRA', 'OriginalNRO', 'OriginalOOA', 'OriginalPFA', 'OriginalPOA', 'AIW_PSO', 'CL_PSO', 'C_PSO', 'HPSO_TVAC', 'LDW_PSO', 'OriginalPSO', 'P_PSO', 'OriginalPSS', 'DevQSA', 'ImprovedQSA', 'LevyQSA', 'OppoQSA', 'OriginalQSA', 'OriginalRIME', 'OriginalRUN', 'GaussianSA', 'OriginalSA', 'SwarmSA', 'DevSARO', 'OriginalSARO', 'DevSBO', 'OriginalSBO', 'DevSCA', 'OriginalSCA', 'QleSCA', 'OriginalSCSO', 'ImprovedSFO', 'OriginalSFO', 'L_SHADE', 'OriginalSHADE', 'OriginalSHIO', 'OriginalSHO', 'ImprovedSLO', 'ModifiedSLO', 'OriginalSLO', 'DevSMA', 'OriginalSMA', 'DevSOA', 'OriginalSOA', 'OriginalSOS', 'DevSPBO', 'OriginalSPBO', 'OriginalSRSR', 'DevSSA', 'OriginalSSA', 'OriginalSSDO', 'OriginalSSO', 'OriginalSSpiderA', 'OriginalSSpiderO', 'OriginalSTO', 'OriginalSeaHO', 'OriginalServalOA', 'OriginalTDO', 'DevTLO', 'ImprovedTLO', 'OriginalTLO', 'OriginalTOA', 'DevTPO', 'OriginalTS', 'OriginalTSA', 'OriginalTSO', 'EnhancedTWO', 'LevyTWO', 'OppoTWO', 'OriginalTWO', 'DevVCS', 'OriginalVCS', 'OriginalWCA', 'OriginalWDO', 'OriginalWHO', 'HI_WOA', 'OriginalWOA', 'OriginalWaOA', 'OriginalWarSO', 'OriginalZOA']

SUPPORTED_REG_OBJECTIVES = {'A10': 'max', 'A20': 'max', 'A30': 'max', 'ACOD': 'max', 'APCC': 'max', 'AR': 'max', 'AR2': 'max', 'CI': 'max', 'COD': 'max', 'COR': 'max', 'COV': 'max', 'CRM': 'min', 'DRV': 'min', 'EC': 'max', 'EVS': 'max', 'GINI': 'min', 'GINI_WIKI': 'min', 'JSD': 'min', 'KGE': 'max', 'MAAPE': 'min', 'MAE': 'min', 'MAPE': 'min', 'MASE': 'min', 'ME': 'min', 'MRB': 'min', 'MRE': 'min', 'MSE': 'min', 'MSLE': 'min', 'MedAE': 'min', 'NNSE': 'max', 'NRMSE': 'min', 'NSE': 'max', 'OI': 'max', 'PCC': 'max', 'PCD': 'max', 'R': 'max', 'R2': 'max', 'R2S': 'max', 'RAE': 'min', 'RMSE': 'min', 'RSE': 'min', 'RSQ': 'max', 'SMAPE': 'min', 'VAF': 'max', 'WI': 'max'}

build_model()

Builds the model architecture and sets the optimizer and loss function based on the task.

Raises:

ValueError – If the task is not recognized.

get_name()

Generate a descriptive name for the MLP model based on the optimizer.

objective_function(solution=None)

Evaluates the fitness function for classification metrics based on the provided solution.

Parameters:

solution (np.ndarray, default=None) – The proposed solution to evaluate.

Returns:

result – The fitness value, representing the loss for the current solution.

Return type:

float

class waveletml.models.mha_wnn.MhaWnnClassifier(size_hidden=10, wavelet_fn='morlet', act_output=None, optim='Adam', optim_params=None, obj_name=None, seed=42, verbose=True, wnn_type: str | Type[BaseCustomWNN] | None = None, lb=None, ub=None, mode='single', n_workers=None, termination=None)

Bases: BaseMhaWnnModel, ClassifierMixin

Metaheuristic-based Wavelet Neural Network (MhaWNN) Classifier.

A classifier that combines wavelet neural networks (WNNs) with metaheuristic optimization techniques to perform supervised classification tasks. The model architecture is based on customizable wavelet functions, and it leverages population-based optimizers to train the weights of the WNN.

Parameters:

• size_hidden (int, optional (default=10)) – Number of hidden neurons in the wavelet neural network.

• wavelet_fn (str, optional (default="morlet")) – Name of the wavelet function to use in hidden layers.

• act_output (callable or None, optional (default=None)) – Activation function for the output layer.

• optim (str, optional (default="Adam")) – Name of the metaheuristic optimizer to use. Must be supported by the Mealpy library.

• optim_params (dict or None, optional (default=None)) – Parameters to configure the optimizer. If None, default parameters are used.

• obj_name (str or None, optional (default=None)) – Name of the classification objective function (metric) to optimize.

• seed (int, optional (default=42)) – Random seed for reproducibility.

• verbose (bool, optional (default=True)) – If True, prints progress during optimization.

• wnn_type (str, type, or None, optional (default=None)) –

  Type of wavelet neural network to use. Accepts:

  • A string name of a WNN class from waveletml.models.custom_wnn

  • A class object that inherits from BaseCustomWNN

  • None to use the default CustomWaveletWeightedLinearNetwork

• lb (float, int, list, tuple, or np.ndarray, optional) – Lower bounds for optimization. If not provided, defaults to -1.0 for each dimension.

• ub (float, int, list, tuple, or np.ndarray, optional) – Upper bounds for optimization. If not provided, defaults to 1.0 for each dimension.

• mode (str, optional (default="single")) – Optimization mode (‘single’, ‘parallel’, etc.).

• n_workers (int or None, optional) – Number of workers for parallel optimization (only used in parallel mode).

• termination (Any, optional) – Termination condition for the optimizer.

size_input

Number of input features in the dataset.

Type:

int

size_output

Number of output classes (1 for binary classification, C for multi-class).

Type:

int

network

Instantiated wavelet neural network model.

Type:

torch.nn.Module

optimizer

Instantiated metaheuristic optimizer.

Type:

Optimizer

classes_

Sorted array of unique class labels.

Type:

np.ndarray

task

Task type, either ‘binary_classification’ or ‘classification’.

Type:

str

minmax

Direction of optimization (‘min’ or ‘max’), based on obj_name.

Type:

str

loss_train

List of training loss values (objective metric) for each epoch.

Type:

list of float

metric_class

Metric class used for evaluation (set to ClassificationMetric).

Type:

callable

fit(X, y)

Trains the classifier using metaheuristic optimization on provided training data.

predict(X)

Predicts class labels for the given input data.

predict_proba(X)

Returns the class probability estimates for the input data.

score(X, y)

Returns the classification accuracy of the model.

evaluate(y_true, y_pred, list_metrics=('AS', 'RS'))

Evaluates model performance using specified classification metrics.

evaluate(y_true, y_pred, list_metrics=('AS', 'RS'))

Return the list of performance metrics on the given test data and labels.

Parameters:

• y_true (array-like of shape (n_samples,) or (n_samples, n_outputs)) – True values for X.

• y_pred (array-like of shape (n_samples,) or (n_samples, n_outputs)) – Predicted values for X.

• list_metrics (list, default=("AS", "RS")) – List of metrics to compute using Permetrics library: https://github.com/thieu1995/permetrics

Returns:

results – A dictionary containing the results of the requested metrics.

Return type:

dict

fit(X, y)

Fits the model to the training data.

Parameters:

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

• y (array-like, shape (n_samples,)) – Target values.

Returns:

self – Returns the instance of the fitted model.

Return type:

MhaWnnClassifier

predict(X)

Predicts the class labels for the provided input data.

Parameters:

X (array-like, shape (n_samples, n_features)) – Input data for prediction.

Returns:

Predicted class labels for each sample.

Return type:

np.ndarray

predict_proba(X)

Computes the probability estimates for each class (for classification tasks only).

Parameters:

X (array-like, shape (n_samples, n_features)) – Input data for which to predict probabilities.

Returns:

Probability predictions for each class.

Return type:

np.ndarray

Raises:

ValueError – If the task is not a classification task.

score(X, y)

Computes the accuracy score of the model based on predictions.

Parameters:

• X (array-like, shape (n_samples, n_features)) – Input data for scoring.

• y (array-like, shape (n_samples,)) – True labels for comparison.

Returns:

Accuracy score of the model.

Return type:

float

class waveletml.models.mha_wnn.MhaWnnRegressor(size_hidden=10, wavelet_fn='morlet', act_output=None, optim='Adam', optim_params=None, obj_name=None, seed=42, verbose=True, wnn_type: str | Type[BaseCustomWNN] | None = None, lb=None, ub=None, mode='single', n_workers=None, termination=None)

Bases: BaseMhaWnnModel, RegressorMixin

Metaheuristic-based Wavelet Neural Network (MhaWNN) Regressor.

A regressor that combines wavelet neural networks (WNNs) with metaheuristic optimization algorithms to solve single- and multi-output regression problems. The model leverages customizable wavelet functions in its architecture, and uses population-based optimizers to train the network parameters.

Parameters:

• size_hidden (int, optional (default=10)) – Number of hidden neurons in the wavelet neural network.

• wavelet_fn (str, optional (default="morlet")) – Name of the wavelet function to use in the hidden layer.

• act_output (callable or None, optional (default=None)) – Activation function to apply at the output layer.

• optim (str, optional (default="Adam")) – Name of the metaheuristic optimizer to use. Must be supported by Mealpy.

• optim_params (dict or None, optional (default=None)) – Additional parameters for the optimizer. If None, defaults are used.

• obj_name (str or None, optional (default=None)) – Name of the objective function (metric) to optimize. Must be supported by permetrics.

• seed (int, optional (default=42)) – Random seed for reproducibility.

• verbose (bool, optional (default=True)) – If True, prints progress and logs during training.

• wnn_type (str, type, or None, optional (default=None)) –

  Specifies the type of WNN to use. Options include:

  • String name of a WNN class defined in waveletml.models.custom_wnn

  • A subclass of BaseCustomWNN

  • None to use the default CustomWaveletWeightedLinearNetwork

• lb (float, int, list, tuple, or np.ndarray, optional) – Lower bounds for the optimizer. Defaults to -1.0 for all weights if not set.

• ub (float, int, list, tuple, or np.ndarray, optional) – Upper bounds for the optimizer. Defaults to 1.0 for all weights if not set.

• mode (str, optional (default="single")) – Optimization mode, e.g., ‘single’ or ‘swarm’, ‘thread’, or ‘process’.

• n_workers (int or None, optional) – Number of parallel workers (if supported by the optimizer).

• termination (any, optional) – Termination criteria for the optimizer.

size_input

Number of input features.

Type:

int

size_output

Number of regression targets (1 for single-output, >1 for multi-output).

Type:

int

network

Instantiated wavelet neural network model.

Type:

torch.nn.Module

optimizer

Configured metaheuristic optimizer.

Type:

Optimizer

task

Type of regression task: ‘regression’ or ‘multi_regression’.

Type:

str

minmax

Optimization direction (‘min’ or ‘max’) based on objective function.

Type:

str

metric_class

Metric class used for evaluation (set to RegressionMetric).

Type:

callable

loss_train

Training losses (metric values) recorded over epochs.

Type:

list of float

fit(X, y)

Trains the model using the specified optimizer on input data X and target y.

predict(X)

Predicts output values for the given input data.

score(X, y)

Returns R² (coefficient of determination) on test data.

evaluate(y_true, y_pred, list_metrics=('AS', 'RS'))

Evaluates regression performance using selected metrics.

evaluate(y_true, y_pred, list_metrics=('AS', 'RS'))

Return the list of performance metrics on the given test data and labels.

Parameters:

• y_true (array-like of shape (n_samples,) or (n_samples, n_outputs)) – True values for X.

• y_pred (array-like of shape (n_samples,) or (n_samples, n_outputs)) – Predicted values for X.

• list_metrics (list, default=("AS", "RS")) – List of metrics to compute using Permetrics library: https://github.com/thieu1995/permetrics

Returns:

results – A dictionary containing the results of the requested metrics.

Return type:

dict

fit(X, y)

Fits the model to the training data.

Parameters:

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

• y (array-like, shape (n_samples,) or (n_samples, n_outputs)) – Target values.

Returns:

self – Returns the instance of the fitted model.

Return type:

MhaWnnRegressor

predict(X)

Predicts the output values for the provided input data.

Parameters:

X (array-like, shape (n_samples, n_features)) – Input data for prediction.

Returns:

Predicted output values for each sample.

Return type:

np.ndarray

score(X, y)

Computes the R2 score of the model based on predictions.

Parameters:

• X (array-like, shape (n_samples, n_features)) – Input data for scoring.

• y (array-like, shape (n_samples,)) – True labels for comparison.

Returns:

R2 score of the model.

Return type:

float