Adds get_variable_inclusion function (#214)

* add get_variable_inclusion function

* add elements to API reference

Author: aloctavodia

docs/api_reference.rst

@@ -13,4 +13,4 @@ methods in the current release of PyMC-BART.
 =============================
 
 .. automodule:: pymc_bart
-   :members: BART, PGBART, plot_pdp, plot_ice, plot_variable_importance, plot_convergence, ContinuousSplitRule, OneHotSplitRule, SubsetSplitRule
+   :members: BART, PGBART, compute_variable_importance, get_variable_inclusion, plot_convergence, plot_ice, plot_pdp, plot_scatter_submodels, plot_variable_importance, plot_variable_inclusion, ContinuousSplitRule, OneHotSplitRule, SubsetSplitRule

pymc_bart/__init__.py

@@ -18,6 +18,7 @@
 from pymc_bart.split_rules import ContinuousSplitRule, OneHotSplitRule, SubsetSplitRule
 from pymc_bart.utils import (
     compute_variable_importance,
+    get_variable_inclusion,
     plot_convergence,
     plot_ice,
     plot_pdp,
@@ -33,6 +34,7 @@
     "OneHotSplitRule",
     "SubsetSplitRule",
     "compute_variable_importance",
+    "get_variable_inclusion",
     "plot_convergence",
     "plot_ice",
     "plot_pdp",

pymc_bart/utils.py

@@ -693,6 +693,50 @@ def _smooth_mean(
     return x_data, y_data
 
 
+def get_variable_inclusion(idata, X, labels=None, to_kulprit=False):
+    """
+    Get the normalized variable inclusion from BART model.
+
+    Parameters
+    ----------
+    idata : InferenceData
+        InferenceData containing a collection of BART_trees in sample_stats group
+    X : npt.NDArray
+        The covariate matrix.
+    labels : Optional[list[str]]
+        List of the names of the covariates. If X is a DataFrame the names of the covariables will
+        be taken from it and this argument will be ignored.
+    to_kulprit : bool
+        If True, the function will return a list of list with the variables names.
+        This list can be passed as a path to Kulprit's project method. Defaults to False.
+    Returns
+    -------
+    VI_norm : npt.NDArray
+        Normalized variable inclusion.
+    labels : list[str]
+        List of the names of the covariates.
+    """
+    VIs = idata["sample_stats"]["variable_inclusion"].mean(("chain", "draw")).values
+    VI_norm = VIs / VIs.sum()
+    idxs = np.argsort(VI_norm)
+
+    indices = idxs[::-1]
+    n_vars = len(indices)
+
+    if hasattr(X, "columns") and hasattr(X, "to_numpy"):
+        labels = X.columns
+
+    if labels is None:
+        labels = np.arange(n_vars).astype(str)
+
+    label_list = labels.to_list()
+
+    if to_kulprit:
+        return [label_list[:idx] for idx in range(n_vars)]
+    else:
+        return VI_norm[indices], label_list
+
+
 def plot_variable_inclusion(idata, X, labels=None, figsize=None, plot_kwargs=None, ax=None):
     """
     Plot normalized variable inclusion from BART model.
@@ -720,26 +764,15 @@ def plot_variable_inclusion(idata, X, labels=None, figsize=None, plot_kwargs=Non
 
     Returns
     -------
-    idxs: indexes of the covariates from higher to lower relative importance
     axes: matplotlib axes
     """
     if plot_kwargs is None:
         plot_kwargs = {}
 
-    VIs = idata["sample_stats"]["variable_inclusion"].mean(("chain", "draw")).values
-    VIs = VIs / VIs.sum()
-    idxs = np.argsort(VIs)
-
-    indices = idxs[::-1]
-    n_vars = len(indices)
-
-    if hasattr(X, "columns") and hasattr(X, "to_numpy"):
-        labels = X.columns
+    VI_norm, labels = get_variable_inclusion(idata, X, labels)
+    n_vars = len(labels)
 
-    if labels is None:
-        labels = np.arange(n_vars).astype(str)
-
-    new_labels = ["+ " + ele if index != 0 else ele for index, ele in enumerate(labels[indices])]
+    new_labels = ["+ " + ele if index != 0 else ele for index, ele in enumerate(labels)]
 
     ticks = np.arange(n_vars, dtype=int)
 
@@ -749,19 +782,18 @@ def plot_variable_inclusion(idata, X, labels=None, figsize=None, plot_kwargs=Non
     if ax is None:
         _, ax = plt.subplots(1, 1, figsize=figsize)
 
+    ax.axhline(1 / n_vars, color="0.5", linestyle="--")
     ax.plot(
-        VIs[indices],
+        VI_norm,
         color=plot_kwargs.get("color", "k"),
         marker=plot_kwargs.get("marker", "o"),
         ls=plot_kwargs.get("ls", "-"),
     )
 
     ax.set_xticks(ticks, new_labels, rotation=plot_kwargs.get("rotation", 0))
-
-    ax.axhline(1 / n_vars, color="0.5", linestyle="--")
     ax.set_ylim(0, 1)
 
-    return idxs, ax
+    return ax
 
 
 def compute_variable_importance(  # noqa: PLR0915 PLR0912