update README.md and docs

README.md

@@ -21,7 +21,10 @@ Available regression models
 - Lasso (`sklearn` may be a better option)
 - Group Lasso, Overlap Group Lasso & Sparse Group Lasso
 - Adaptive versions of Lasso, Group Lasso, Overlap Group Lasso & Sparse Group Lasso
-- Best subset selection, L1L0 & L2L0 (we recommend using `gurobi` for performance)
+- Best subset selection, ridged best subset, L0, L1L0 & L2L0
+  (`gurobi` recommended for performance)
+- Best group selection, ridged best group selection,  grouped L0, grouped L2L0
+  (`gurobi` recommended for performance)
 
 Installation
 ------------

sparselm/model/miqp/best_subset.py

@@ -122,7 +122,7 @@ def _gen_hierarchy_constraints(self):
 
 
 class RidgedBestSubsetSelection(BestSubsetSelection):
-    """Best subset selection estimaor with ridge regularization."""
+    """MIQP  Best subset selection estimator with ridge regularization."""
 
     def __init__(self, sparse_bound, alpha=1.0, big_M=1000, hierarchy=None,
                  ignore_psd_check=True, fit_intercept=False, normalize=False,
@@ -144,7 +144,7 @@ def __init__(self, sparse_bound, alpha=1.0, big_M=1000, hierarchy=None,
                 coefficient 0 depends on 1, and 2; 1 depends on 0, and 2 has no
                 dependence.
             ignore_psd_check (bool):
-                Wether to ignore cvxpy's PSD checks  of matrix used in quadratic
+                Whether to ignore cvxpy's PSD checks  of matrix used in quadratic
                 form. Default is True to avoid raising errors for poorly
                 conditioned matrices. But if you want to be strict set to False.
             fit_intercept (bool):
@@ -195,7 +195,7 @@ def _gen_objective(self, X, y):
 
 
 class BestGroupSelection(BestSubsetSelection):
-    """Best group selection estimator."""
+    """MIQP Best group selection estimator."""
 
     def __init__(self, groups, sparse_bound, big_M=1000, hierarchy=None,
                  ignore_psd_check=True, fit_intercept=False, normalize=False,

sparselm/model/miqp/regularized_l0.py

@@ -20,7 +20,7 @@
 
 
 class RegularizedL0(CVXEstimator):
-    """Implementation of l0 regularized estimator."""
+    """Implementation of MIQP l0 regularized estimator."""
 
     def __init__(self, alpha=1.0, big_M=1000, hierarchy=None, ignore_psd_check=True,
                  fit_intercept=False, normalize=False, copy_X=True, warm_start=False,
@@ -315,6 +315,8 @@ class L2L0(MixedL0):
     Estimator with L2L0 regularization solved with mixed integer programming
     proposed by Peichen Zhong.
 
+    https://arxiv.org/abs/2204.13789
+
     Regularized model is:
         ||X * Beta - y||^2 + alpha * l0_ratio * ||Beta||_0
                            + alpha * (1 - l0_ratio) * ||Beta||^2_2