- Add metrics AUC, Roc curve

evo_science/entities/metrics/__init__.py

@@ -1,11 +1,14 @@
+from .accuracy import Accuracy
+from .auc import AUC
 from .base_metric import BaseMetric
-from .slope import Slope
-from .rmse import RMSE
-from .mae import MAE
 from .error_ave import ErrorAve
-from .r_squared import RSquared
 from .error_ave import ErrorAve
 from .error_std import ErrorStd
-from .accuracy import Accuracy
-from .precision import Precision
 from .f1_score import F1Score
+from .mae import MAE
+from .precision import Precision
+from .r_squared import RSquared
+from .recall import Recall
+from .rmse import RMSE
+from .roc_curve import RocCurve
+from .slope import Slope

evo_science/entities/metrics/auc.py

@@ -0,0 +1,36 @@
+import matplotlib.pyplot as plt
+from sklearn import metrics
+
+from evo_science.entities.metrics.base_binary_classify_metric import (
+    BaseBinaryClassifyMetric,
+)
+
+
+class AUC(BaseBinaryClassifyMetric):
+    name = "Area Under the Curve"
+
+    def _on_init(self, threshold=0.5, plot=False, **kwargs):
+        self.threshold = threshold
+        self.plot = plot
+
+    def _calculate_np(self, y_true, y_pred):
+        auc = metrics.roc_auc_score(y_true, y_pred)
+        if self.plot:
+            fpr, tpr, _ = metrics.roc_curve(y_true, y_pred)
+            plt.figure()
+            plt.plot(
+                fpr,
+                tpr,
+                color="darkorange",
+                lw=2,
+                label="ROC curve (area = %0.2f)" % auc,
+            )
+            plt.plot([0, 1], [0, 1], color="navy", lw=2, linestyle="--")
+            plt.xlim([0.0, 1.0])
+            plt.ylim([0.0, 1.05])
+            plt.xlabel("False Positive Rate")
+            plt.ylabel("True Positive Rate")
+            plt.title("Receiver Operating Characteristic Curve")
+            plt.legend(loc="lower right")
+            plt.show()
+        return auc

evo_science/entities/metrics/recall.py

@@ -4,6 +4,7 @@
 
 
 class Recall(BaseBinaryClassifyMetric):
+    name = "Recall"
 
     def _calculate_np(self, y_true: np.array, y_pred: np.array):
         y_pred = self._binary_threshold(y_pred)

evo_science/entities/metrics/roc_curve.py

@@ -0,0 +1,26 @@
+import matplotlib.pyplot as plt
+import numpy as np
+from sklearn.metrics import roc_curve
+
+from evo_science.entities.metrics.base_binary_classify_metric import (
+    BaseBinaryClassifyMetric,
+)
+
+
+class RocCurve(BaseBinaryClassifyMetric):
+    name = "Roc Curve"
+
+    def _on_init(self, threshold=0.5, plot=False, **kwargs):
+        self.threshold = threshold
+        self.plot = plot
+
+    def _calculate_np(self, y_true: np.array, y_pred: np.array):
+        fpr, tpr, thresholds = roc_curve(y_true, y_pred)
+        if self.plot:
+            plt.plot(fpr, tpr)
+            plt.xlabel("False Positive Rate")
+            plt.ylabel("True Positive Rate")
+            plt.title("ROC Curve")
+            plt.show()
+
+        return fpr.mean(), tpr.mean(), thresholds.mean()

examples/titanic_survival_prediction/linear_regression_model.py

@@ -49,7 +49,18 @@ def example_lr_model():
 
     model.fit(x=x, y=y)
     model.evaluate(
-        x=x, y=y, metrics=[Slope, ErrorStd, Accuracy(threshold=0.5), Precision, F1Score]
+        x=x,
+        y=y,
+        metrics=[
+            Slope,
+            ErrorStd,
+            Accuracy(threshold=0.5),
+            Precision,
+            Recall,
+            F1Score,
+            RocCurve,
+            AUC,
+        ],
     )
     model.calculate_coefficients(x=x)