Merge pull request #20 from Leona-LYT/main

modified `sen_idx` and add _sklearn_mixin.py

Author: statmlben

doc/source/examples/.ipynb_checkpoints/FairSVM-checkpoint.ipynb

@@ -56,7 +56,7 @@
     "X = scaler.fit_transform(X)\n",
     "\n",
     "## we take the first column of X as sensetive features, and tol is 0.1\n",
-    "X_sen = X[:,0]\n",
+    "sen_idx = [0]\n",
     "tol_sen = 0.1"
    ]
   },
@@ -102,7 +102,7 @@
     "\n",
     "fclf = plqERM_Ridge(loss={'name': 'svm'},\n",
     "                   constraint=[{'name': 'fair',\n",
-    "                                'X_sen': X_sen,\n",
+    "                                'sen_idx': sen_idx,\n",
     "                                'tol_sen': tol_sen}],\n",
     "                   C=1.0,\n",
     "                   max_iter=50000)\n",
@@ -142,8 +142,8 @@
     "svm_perf = len(y[score*y > 0])/n\n",
     "fsvm_perf = len(y[fscore*y > 0])/n\n",
     "\n",
-    "svm_corr = score.dot(X_sen) / n\n",
-    "fsvm_corr = fscore.dot(X_sen) / n\n",
+    "svm_corr = score.dot(X[:,sen_idx]) / n\n",
+    "fsvm_corr = fscore.dot(X[:,sen_idx]) / n\n",
     "\n",
     "# Create a pandas DataFrame to store the results\n",
     "results = pd.DataFrame({\n",

doc/source/examples/FairSVM.ipynb

@@ -35,10 +35,10 @@
    ]
   },
   {
-   "metadata": {},
    "cell_type": "markdown",
-   "source": "",
-   "id": "7bf9272115591bbf"
+   "id": "7bf9272115591bbf",
+   "metadata": {},
+   "source": []
   },
   {
    "cell_type": "code",
@@ -61,7 +61,7 @@
     "X = scaler.fit_transform(X)\n",
     "\n",
     "## we take the first column of X as sensetive features, and tol is 0.1\n",
-    "X_sen = X[:,0]\n",
+    "sen_idx = [0]\n",
     "tol_sen = 0.1"
    ]
   },
@@ -107,7 +107,7 @@
     "\n",
     "fclf = plqERM_Ridge(loss={'name': 'svm'},\n",
     "                   constraint=[{'name': 'fair',\n",
-    "                                'X_sen': X_sen,\n",
+    "                                'sen_idx': sen_idx,\n",
     "                                'tol_sen': tol_sen}],\n",
     "                   C=1.0,\n",
     "                   max_iter=50000)\n",
@@ -205,7 +205,7 @@
    ]
   }
  ],
-  "metadata": {
+ "metadata": {
   "colab": {
    "provenance": []
   },
@@ -219,4 +219,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}

doc/source/examples/Path_solution.ipynb

@@ -176,10 +176,10 @@
       "outputs": [],
       "source": [
         "## we take the first column of X as sensetive features, and tol is 0.1\n",
-        "X_sen = X[:,0]\n",
+        "sen_idx = [0]\n",
         "tol_sen = 0.1\n",
         "# define constraint\n",
-        "constraint=[{'name': 'fair', 'X_sen': X_sen,'tol_sen': tol_sen}]\n",
+        "constraint=[{'name': 'fair', 'sen_idx': sen_idx,'tol_sen': tol_sen}]\n",
         "# define loss function\n",
         "loss = {'name': 'SVM'}\n",
         "# define value of Cs\n",

doc/source/getting_started.rst

@@ -126,7 +126,7 @@ With `rehline`, you can easily add a `fairness constraint` to your ERM.
     from scipy.stats import pearsonr
     # Define a Huber estimator with fairness constraint
     clf = plqERM_Ridge(loss={'name': 'huber', 'tau': 0.5},
-                       constraint=[{'name': 'fair', 'X_sen': X_train[:, 0], 'tol_sen': 0.1}], 
+                       constraint=[{'name': 'fair', 'sen_idx': [0], 'tol_sen': 0.1}], 
                        C=1.0,
                        max_iter=10000)
     clf.fit(X=X_train, y=y_train)

doc/source/tutorials/constraint.rst

@@ -28,9 +28,9 @@ Usage
 
  * - **fair**
    - | ``name``: 'fair' or 'fairness'
-     | ``X_sen``: 2d array [n x p] for sensitive attributes
+     | ``sen_idx``: a list contains column indices for sensitive attributes
      | ``tol_sen``: 1d array [p] of tolerance for fairness
-   - | ``constraint=[{'name': 'fair', 'X_sen': X_sen, 'tol_sen': tol_sen}]``
+   - | ``constraint=[{'name': 'fair', 'sen_idx': sen_idx, 'tol_sen': tol_sen}]``
 
  * - **custom**
    - | ``name``: 'custom'