add benchmark

Author: statmlben

python/benchmark/benchmark_FairSVM/datasets/classification_sim.py

@@ -0,0 +1,41 @@
+from benchopt import BaseDataset, safe_import_context
+from sklearn.preprocessing import StandardScaler
+import numpy as np
+with safe_import_context() as import_ctx:
+    from benchopt.datasets import make_correlated_data
+
+
+class Dataset(BaseDataset):
+
+    name = "Classification_sim"
+
+    parameters = {
+        'n_samples, n_features': [
+            (100, 10),
+            (300, 10),
+            (500, 10),
+            (700, 10),
+            (1000, 10)]
+    }
+
+    def __init__(self, n_samples=10, n_features=50, random_state=42):
+        self.n_samples = n_samples
+        self.n_features = n_features
+        self.random_state = random_state
+
+    def get_data(self):
+        X, y, _ = make_correlated_data(
+            n_samples=self.n_samples, n_features=self.n_features,
+            random_state=self.random_state
+        )
+        y = 2*(y > 0) - 1
+
+        scaler = StandardScaler()
+        X = scaler.fit_transform(X)
+
+        n, d = X.shape
+        x_sensitive = X[:,np.random.randint(d)]
+        x_sensitive = x_sensitive - np.mean(x_sensitive)
+        data = dict(X=X, y=y, Z=x_sensitive)
+
+        return self.n_features, data

python/benchmark/benchmark_FairSVM/datasets/reg_sim.py

@@ -0,0 +1,38 @@
+from benchopt import BaseDataset, safe_import_context
+from sklearn.preprocessing import StandardScaler
+import numpy as np
+with safe_import_context() as import_ctx:
+    from benchopt.datasets import make_correlated_data
+
+
+class Dataset(BaseDataset):
+
+    name = "reg_sim"
+
+    parameters = {
+        'n_samples, n_features': [
+            (100, 10),
+            (300, 10),
+            (500, 10),
+            (700, 10),
+            (1000, 10)]
+    }
+
+    def __init__(self, n_samples=10, n_features=50, random_state=42):
+        self.n_samples = n_samples
+        self.n_features = n_features
+        self.random_state = random_state
+
+    def get_data(self):
+        X, y, _ = make_correlated_data(
+            n_samples=self.n_samples, n_features=self.n_features,
+            random_state=self.random_state
+        )
+
+        scaler = StandardScaler()
+        X = scaler.fit_transform(X)
+
+        n, d = X.shape
+        data = dict(X=X, y=y, q=0.1)
+
+        return self.n_features, data