Liking cljdoc? Tell your friends :D

scicloj.metamorph.ml.loss

Simple loss functions.

Simple loss functions.
raw docstring

aucclj

(auc predictions labels)

Calculates area under the ROC curve for binary classification.

predictions - Sequence of predicted scores/probabilities (numeric) labels - Sequence of binary labels (must be 0 or 1)

Returns the AUC score as a double in [0, 1]. Values above 0.5 indicate the model performs better than random guessing. AUC of 1.0 is perfect classification.

Uses the Mann-Whitney U statistic formula from R's 'mlr' package: (sum(r[i]) - n.pos * (n.pos + 1) / 2) / (n.pos * n.neg)

Both inputs must have equal length, and labels must contain only 0 and 1 values.

Reference: https://github.com/mlr-org/mlr/blob/main/R/measures.R

See also: classification-accuracy

Calculates area under the ROC curve for binary classification.

`predictions` - Sequence of predicted scores/probabilities (numeric)
`labels` - Sequence of binary labels (must be 0 or 1)

Returns the AUC score as a double in [0, 1]. Values above 0.5 indicate the
model performs better than random guessing. AUC of 1.0 is perfect classification.

Uses the Mann-Whitney U statistic formula from R's 'mlr' package:
`(sum(r[i]) - n.pos * (n.pos + 1) / 2) / (n.pos * n.neg)`

Both inputs must have equal length, and labels must contain only 0 and 1 values.

Reference: https://github.com/mlr-org/mlr/blob/main/R/measures.R

See also: `classification-accuracy`
sourceraw docstring

classification-accuracyclj

(classification-accuracy lhs rhs)

Calculates classification accuracy as the proportion of correct predictions.

lhs - Sequence of predicted class labels (without categorical map metadata) rhs - Sequence of actual/ground truth class labels (without categorical map metadata)

Returns accuracy as a double in [0, 1] where 1.0 is perfect classification. Computes the ratio of correct predictions to total predictions. Both inputs must have the same length and be either both numeric or both non-numeric.

Note: Categorical maps should be removed from columns before calling this function. See the link in error messages for details on categorical map handling.

See also: classification-loss, scicloj.metamorph.ml.metrics/accuracy

Calculates classification accuracy as the proportion of correct predictions.

`lhs` - Sequence of predicted class labels (without categorical map metadata)
`rhs` - Sequence of actual/ground truth class labels (without categorical map metadata)

Returns accuracy as a double in [0, 1] where 1.0 is perfect classification.
Computes the ratio of correct predictions to total predictions. Both inputs
must have the same length and be either both numeric or both non-numeric.

Note: Categorical maps should be removed from columns before calling this function.
See the link in error messages for details on categorical map handling.

See also: `classification-loss`, `scicloj.metamorph.ml.metrics/accuracy`
sourceraw docstring

classification-lossclj

(classification-loss lhs rhs)

Calculates classification loss as the proportion of incorrect predictions.

lhs - Sequence of predicted class labels (without categorical map metadata) rhs - Sequence of actual/ground truth class labels (without categorical map metadata)

Returns classification error rate as a double in [0, 1] where 0.0 is perfect classification. Computed as 1.0 minus classification accuracy. Lower values indicate better model performance.

Commonly used as the default loss function for classification in evaluate-pipelines.

See also: classification-accuracy, scicloj.metamorph.ml.metrics/error-rate

Calculates classification loss as the proportion of incorrect predictions.

`lhs` - Sequence of predicted class labels (without categorical map metadata)
`rhs` - Sequence of actual/ground truth class labels (without categorical map metadata)

Returns classification error rate as a double in [0, 1] where 0.0 is perfect
classification. Computed as 1.0 minus classification accuracy. Lower values
indicate better model performance.

Commonly used as the default loss function for classification in `evaluate-pipelines`.

See also: `classification-accuracy`, `scicloj.metamorph.ml.metrics/error-rate`
sourceraw docstring

maeclj

(mae predictions labels)

Calculates mean absolute error between predictions and labels.

predictions - Sequence of predicted values labels - Sequence of actual/ground truth values

Returns the average of absolute differences between predictions and labels. MAE is more robust to outliers than MSE/RMSE and shares the same units as the target variable. Lower values indicate better model fit.

Commonly used as the default loss function for regression in evaluate-pipelines.

See also: mse, rmse

Calculates mean absolute error between predictions and labels.

`predictions` - Sequence of predicted values
`labels` - Sequence of actual/ground truth values

Returns the average of absolute differences between predictions and labels.
MAE is more robust to outliers than MSE/RMSE and shares the same units as the
target variable. Lower values indicate better model fit.

Commonly used as the default loss function for regression in `evaluate-pipelines`.

See also: `mse`, `rmse`
sourceraw docstring

mseclj

(mse predictions labels)

Calculates mean squared error between predictions and labels.

predictions - Sequence of predicted values labels - Sequence of actual/ground truth values

Returns the average of squared differences between predictions and labels. Lower values indicate better model fit. Commonly used for regression models.

See also: rmse, mae

Calculates mean squared error between predictions and labels.

`predictions` - Sequence of predicted values
`labels` - Sequence of actual/ground truth values

Returns the average of squared differences between predictions and labels.
Lower values indicate better model fit. Commonly used for regression models.

See also: `rmse`, `mae`
sourceraw docstring

rmseclj

(rmse predictions labels)

Calculates root mean squared error between predictions and labels.

predictions - Sequence of predicted values labels - Sequence of actual/ground truth values

Returns the square root of mean squared error. RMSE is in the same units as the target variable, making it more interpretable than MSE. Lower values indicate better model fit.

See also: mse, mae

Calculates root mean squared error between predictions and labels.

`predictions` - Sequence of predicted values
`labels` - Sequence of actual/ground truth values

Returns the square root of mean squared error. RMSE is in the same units as
the target variable, making it more interpretable than MSE. Lower values
indicate better model fit.

See also: `mse`, `mae`
sourceraw docstring

cljdoc builds & hosts documentation for Clojure/Script libraries

Keyboard shortcuts
Ctrl+kJump to recent docs
Move to previous article
Move to next article
Ctrl+/Jump to the search field
× close