0.10.64

plot_confusion_matrix layout improved
fun_control_init accepts weights arguments
new get_metric_sgn function

Author: bartzbeielstein

pyproject.toml

@@ -7,7 +7,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "spotPython"
-version = "0.10.63"
+version = "0.10.64"
 authors = [
   { name="T. Bartz-Beielstein", email="tbb@bartzundbartz.de" }
 ]

src/spotPython/plot/validation.py

@@ -172,7 +172,15 @@ def plot_roc_from_dataframes(
 
 
 def plot_confusion_matrix(
-    model=None, fun_control=None, df=None, title=None, target_names=None, y_true_name=None, y_pred_name=None, show=False
+    model=None,
+    fun_control=None,
+    df=None,
+    title=None,
+    target_names=None,
+    y_true_name=None,
+    y_pred_name=None,
+    show=False,
+    ax=None,
 ):
     """
     Plotting a confusion matrix. If a model and the fun_control dictionary are passed,
@@ -197,6 +205,8 @@ def plot_confusion_matrix(
             Name of the column with the predicted values if a dataframe is specified. Defaults to None.
         show (bool, optional):
             If True, the plot is shown. Defaults to False.
+        ax (matplotlib.axes._subplots.AxesSubplot, optional):
+            Axes to plot the confusion matrix. Defaults to None.
 
     Returns:
         (NoneType): None
@@ -212,8 +222,9 @@ def plot_confusion_matrix(
         X_test, y_true = get_Xy_from_df(fun_control["test"], fun_control["target_column"])
         model.fit(X_train, y_train)
         y_pred = model.predict(X_test)
-    fig, ax = plt.subplots(figsize=(10, 5))
-    ConfusionMatrixDisplay.from_predictions(y_true=y_true, y_pred=y_pred, ax=ax)
+    if ax is None:
+        fig, ax = plt.subplots(figsize=(10, 5))
+    ConfusionMatrixDisplay.from_predictions(y_true=y_true, y_pred=y_pred, ax=ax, colorbar=False)
     if target_names is not None:
         ax.xaxis.set_ticklabels(target_names)
         ax.yaxis.set_ticklabels(target_names)

src/spotPython/utils/init.py

@@ -55,6 +55,8 @@ def fun_control_init(
     var_name=None,
     var_type=["num"],
     verbosity=0,
+    weights=1.0,
+    weight_coeff=0.0,
 ):
     """Initialize fun_control dictionary.
 
@@ -171,6 +173,12 @@ def fun_control_init(
         verbosity (int):
             The verbosity level. Determines print output to console. Higher values
             result in more output. Default is 0.
+        weights (float):
+            The weight coefficient of the objective function. Positive values mean minimization.
+            If set to -1, scores that are better when maximized will be minimized, e.g, accuracy.
+            Can be an array, so that different weights can be used for different (multiple) objectives.
+        weight_coeff (float):
+            Determines how to weight older measures. Default is 1.0. Used in the OML algorithm eval_oml.py.
 
     Returns:
         fun_control (dict):
@@ -327,7 +335,8 @@ def fun_control_init(
         "var_name": var_name,
         "var_type": var_type,
         "verbosity": verbosity,
-        "weights": 1.0,
+        "weights": weights,
+        "weight_coeff": weight_coeff,
     }
     # lower = X_reshape(lower)
     # fun_control.update({"lower": lower})

src/spotPython/utils/metrics.py

@@ -128,3 +128,52 @@ def mapk_scorer(estimator, X, y):
     y_pred = estimator.predict_proba(X)
     score = mapk_score(y, y_pred, k=3)
     return score
+
+
+def get_metric_sign(metric_name):
+    """Returns the sign of a metric.
+
+    Args:
+        metric_name (str):
+            The name of the metric. Can be one of the following:
+                - "accuracy_score"
+                - "cohen_kappa_score"
+                - "f1_score"
+                - "hamming_loss"
+                - "hinge_loss"
+                -"jaccard_score"
+                - "matthews_corrcoef"
+                - "precision_score"
+                - "recall_score"
+                - "roc_auc_score"
+                - "zero_one_loss"
+
+    Returns:
+        sign (float): The sign of the metric. -1 for max, +1 for min.
+
+    Raises:
+        ValueError: If the metric is not found.
+
+    Examples:
+        >>> from spotPython.metrics import get_metric_sign
+        >>> get_metric_sign("accuracy_score")
+        -1
+        >>> get_metric_sign("hamming_loss")
+        +1
+
+    """
+    if metric_name in [
+        "accuracy_score",
+        "cohen_kappa_score",
+        "f1_score",
+        "jaccard_score",
+        "matthews_corrcoef",
+        "precision_score",
+        "recall_score",
+        "roc_auc_score",
+    ]:
+        return -1
+    elif metric_name in ["hamming_loss", "hinge_loss", "zero_one_loss"]:
+        return +1
+    else:
+        raise ValueError(f"Metric '{metric_name}' not found.")

test/test_metrics_sign.py

@@ -0,0 +1,19 @@
+from spotPython.utils.metrics import get_metric_sign
+import pytest
+
+def test_get_metric_sign():
+    # Test for accuracy_score
+    assert get_metric_sign("accuracy_score") == -1
+
+    # Test for hamming_loss
+    assert get_metric_sign("hamming_loss") == +1
+
+    # Test for f1_score
+    assert get_metric_sign("f1_score") == -1
+
+    # Test for roc_auc_score
+    assert get_metric_sign("roc_auc_score") == -1
+
+    # Test for unknown metric
+    with pytest.raises(ValueError):
+        get_metric_sign("unknown_metric")