added weighted centering

src/elexsolver/OLSRegressionSolver.py

@@ -82,6 +82,9 @@ def _compute_normal_equations(
         # lambda_I is the regularization matrix
         return np.linalg.inv(R.T @ R + lambda_I) @ R.T @ Q.T
 
+    def _normalize_weights(self, weights):
+        return weights / weights.sum()
+
     def fit(
         self,
         x: np.ndarray,
@@ -105,7 +108,7 @@ def fit(
 
         # normalize weights and turn into diagional matrix
         # square root because will be squared when R^T R happens later
-        L = np.diag(np.sqrt(weights.flatten() / weights.sum()))
+        L = np.diag(np.sqrt(self._normalize_weights(weights.flatten())))
 
         # if normal equations are provided then use those, otherwise compute them
         # in the bootstrap setting we can now pass in the normal equations and can
@@ -123,7 +126,7 @@ def fit(
         # compute coefficients: (X^T X)^{-1} X^T y
         self.coefficients = self.normal_eqs @ L @ y
 
-    def residuals(self, y: np.ndarray, y_hat: np.ndarray, loo: bool = True, center: bool = True) -> np.ndarray:
+    def residuals(self, y: np.ndarray, y_hat: np.ndarray, weights: np.ndarray, loo: bool = True, center: bool = True) -> np.ndarray:
         """
         Computes residuals for the model
         """
@@ -136,8 +139,8 @@ def residuals(self, y: np.ndarray, y_hat: np.ndarray, loo: bool = True, center:
         if loo:
             residuals /= (1 - self.hat_vals).reshape(-1, 1)
 
-        # centering removes the column mean
+        # centering removes the column weighted mean
         if center:
-            residuals -= np.mean(residuals, axis=0)
+            residuals -= np.sum(residuals * self._normalize_weights(weights), axis=0)
 
         return residuals