ENH: Improving GP models

scripts/dwi_gp_estimation_error_analysis.py

@@ -180,6 +180,10 @@ def main() -> None:
         kernel=SphericalKriging(a=a, lambda_s=lambda_s),
         alpha=alpha,
         optimizer=None,
+        # optimizer="Nelder-Mead",
+        # disp=True,
+        # ftol=1,
+        # max_iter=2e5,
     )
 
     # Use Scikit-learn cross validation

src/eddymotion/model/_dipy.py

@@ -92,8 +92,8 @@ class GaussianProcessModel(ReconstModel):
     def __init__(
         self,
         kernel_model: str = "spherical",
-        lambda_s: float = 2.0,
-        a: float = 0.1,
+        beta_l: float = 2.0,
+        beta_a: float = 0.1,
         sigma_sq: float = 1.0,
         *args,
         **kwargs,
@@ -129,8 +129,8 @@ def __init__(
 
         KernelType = SphericalKriging if kernel_model == "spherical" else ExponentialKriging
         self.kernel = KernelType(
-            a=a,
-            lambda_s=lambda_s,
+            beta_a=beta_a,
+            beta_l=beta_l,
         )
 
     def fit(