alpbench.util.pytorch_tabnet.metrics

Functions

UnsupervisedLoss(y_pred, embedded_x, obf_vars)	Implements unsupervised loss function.
UnsupervisedLossNumpy(y_pred, embedded_x, ...)
check_metrics(metrics)	Check if custom metrics are provided.

Classes

AUC()	AUC.
Accuracy()	Accuracy.
BalancedAccuracy()	Balanced Accuracy.
LogLoss()	LogLoss.
MAE()	Mean Absolute Error.
MSE()	Mean Squared Error.
Metric()
MetricContainer(metric_names[, prefix])	Container holding a list of metrics.
RMSE()	Root Mean Squared Error.
RMSLE()	Root Mean squared logarithmic error regression loss.
UnsupMetricContainer(metric_names[, prefix])	Container holding a list of metrics.
UnsupervisedMetric()	Unsupervised metric
UnsupervisedNumpyMetric()	Unsupervised metric

class alpbench.util.pytorch_tabnet.metrics.AUC

Bases: Metric

AUC.

__call__(y_true, y_score)

Compute AUC of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

AUC of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.Accuracy

Bases: Metric

Accuracy.

__call__(y_true, y_score)

Compute Accuracy of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

Accuracy of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.BalancedAccuracy

Bases: Metric

Balanced Accuracy.

__call__(y_true, y_score)

Compute Accuracy of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

Accuracy of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.LogLoss

Bases: Metric

LogLoss.

__call__(y_true, y_score)

Compute LogLoss of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

LogLoss of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.MAE

Bases: Metric

Mean Absolute Error.

__call__(y_true, y_score)

Compute MAE (Mean Absolute Error) of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

MAE of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.MSE

Bases: Metric

Mean Squared Error.

__call__(y_true, y_score)

Compute MSE (Mean Squared Error) of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

MSE of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.Metric

Bases: object

__call__(y_true, y_pred)

classmethod get_metrics_by_names(names)

Get list of metric classes.

Parameters:

• cls (Metric) – Metric class.

• names (list) – List of metric names.

Returns:

metrics – List of metric classes.

Return type:

list

class alpbench.util.pytorch_tabnet.metrics.MetricContainer(metric_names, prefix='')

Bases: object

Container holding a list of metrics.

Parameters:

• metric_names (list of str) – List of metric names.

• prefix (str) – Prefix of metric names.

__call__(y_true, y_pred)

Compute all metrics and store into a dict.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_pred (np.ndarray) – Score matrix or vector

Returns:

Dict of metrics ({metric_name: metric_value}).

Return type:

dict

metric_names: list[str]

prefix: str = ''

class alpbench.util.pytorch_tabnet.metrics.RMSE

Bases: Metric

Root Mean Squared Error.

__call__(y_true, y_score)

Compute RMSE (Root Mean Squared Error) of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

RMSE of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.RMSLE

Bases: Metric

Root Mean squared logarithmic error regression loss. Scikit-implementation: https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_squared_log_error.html Note: In order to avoid error, negative predictions are clipped to 0. This means that you should clip negative predictions manually after calling predict.

__call__(y_true, y_score)

Compute RMSLE of predictions.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_score (np.ndarray) – Score matrix or vector

Returns:

RMSLE of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.UnsupMetricContainer(metric_names, prefix='')

Bases: object

Container holding a list of metrics.

Parameters:

• y_pred (torch.Tensor or np.array) – Reconstructed prediction (with embeddings)

• embedded_x (torch.Tensor) – Original input embedded by network

• obf_vars (torch.Tensor) – Binary mask for obfuscated variables. 1 means the variables was obfuscated so reconstruction is based on this.

__call__(y_pred, embedded_x, obf_vars)

Compute all metrics and store into a dict.

Parameters:

• y_true (np.ndarray) – Target matrix or vector

• y_pred (np.ndarray) – Score matrix or vector

Returns:

Dict of metrics ({metric_name: metric_value}).

Return type:

dict

metric_names: list[str]

prefix: str = ''

class alpbench.util.pytorch_tabnet.metrics.UnsupervisedMetric

Bases: Metric

Unsupervised metric

__call__(y_pred, embedded_x, obf_vars)

Compute MSE (Mean Squared Error) of predictions.

Parameters:

• y_pred (torch.Tensor or np.array) – Reconstructed prediction (with embeddings)

• embedded_x (torch.Tensor) – Original input embedded by network

• obf_vars (torch.Tensor) – Binary mask for obfuscated variables. 1 means the variables was obfuscated so reconstruction is based on this.

Returns:

MSE of predictions vs targets.

Return type:

float

class alpbench.util.pytorch_tabnet.metrics.UnsupervisedNumpyMetric

Bases: Metric

Unsupervised metric

__call__(y_pred, embedded_x, obf_vars)

Compute MSE (Mean Squared Error) of predictions.

Parameters:

• y_pred (torch.Tensor or np.array) – Reconstructed prediction (with embeddings)

• embedded_x (torch.Tensor) – Original input embedded by network

• obf_vars (torch.Tensor) – Binary mask for obfuscated variables. 1 means the variables was obfuscated so reconstruction is based on this.

Returns:

MSE of predictions vs targets.

Return type:

float

alpbench.util.pytorch_tabnet.metrics.UnsupervisedLoss(y_pred, embedded_x, obf_vars, eps=1e-09)

Implements unsupervised loss function. This differs from orginal paper as it’s scaled to be batch size independent and number of features reconstructed independent (by taking the mean)

Parameters:

• y_pred (torch.Tensor or np.array) – Reconstructed prediction (with embeddings)

• embedded_x (torch.Tensor) – Original input embedded by network

• obf_vars (torch.Tensor) – Binary mask for obfuscated variables. 1 means the variable was obfuscated so reconstruction is based on this.

• eps (float) – A small floating point to avoid ZeroDivisionError This can happen in degenerated case when a feature has only one value

Returns:

loss – Unsupervised loss, average value over batch samples.

Return type:

torch float

alpbench.util.pytorch_tabnet.metrics.UnsupervisedLossNumpy(y_pred, embedded_x, obf_vars, eps=1e-09)

alpbench.util.pytorch_tabnet.metrics.check_metrics(metrics)

Check if custom metrics are provided.

Parameters:

metrics (list of str or classes) – List with built-in metrics (str) or custom metrics (classes).

Returns:

val_metrics – List of metric names.

Return type:

list of str