scikit-optimize · holgern · Feb 28, 2020 · Feb 28, 2020 · Feb 28, 2020 · Feb 28, 2020
diff --git a/doc/conf.py b/doc/conf.py
@@ -308,7 +308,8 @@ def __call__(self, directory):
 # thumbnails for the front page of the scikit-learn home page.
 # key: first image in set
 # values: (number of plot in set, height of thumbnail)
-carousel_thumbs = {'sphx_glr_plot_ask-and-tell_002.png': 600,
+carousel_thumbs = {'sphx_glr_sklearn-gridsearchcv-replacement_001.png': 600,
+                   'sphx_glr_plot_ask-and-tell_002.png': 600,
                    'sphx_glr_bayesian-optimization_004.png': 600,
                    'sphx_glr_strategy-comparison_002.png': 600,
                    'sphx_glr_visualizing-results_008.png': 600}

diff --git a/doc/templates/index.html b/doc/templates/index.html
@@ -35,6 +35,7 @@ <h4 class="sk-landing-subheader text-dark font-italic mb-3">Sequential model-bas
         </div>
         <div class="overflow-hidden mx-2 text-center flex-fill">
           <a href="auto_examples/sklearn-gridsearchcv-replacement.html"  aria-label="sklearn-gridsearchcv-replacement">
+            <img src="_images/sphx_glr_sklearn-gridsearchcv-replacement_001.png" class="sk-index-img" alt="Scikit-learn hyperparameter search wrapper">
           </a>
         </div>
       </div>

diff --git a/doc/whats_new/v0.8.rst b/doc/whats_new/v0.8.rst
@@ -37,6 +37,10 @@ Version 0.8.0
 - |Enhancement| Add `is_constant`  property to dimension and
   `n_constant_dimensions` property to Space
   :pr:`883` by :user:`Holger Nahrstaedt <holgern>`
+- |Enhancement| Skip constant dimensions for plot_objective and
+  plot_evaluations
+  to allow plots using BayesSearchCV
+  :pr:`888` by :user:`Holger Nahrstaedt <holgern>`
 
 :mod:`skopt.utils`
 ------------------

diff --git a/examples/sklearn-gridsearchcv-replacement.py b/examples/sklearn-gridsearchcv-replacement.py
@@ -35,6 +35,8 @@
 """
 print(__doc__)
 import numpy as np
+np.random.seed(123)
+import matplotlib.pyplot as plt
 
 #############################################################################
 # Minimal example
@@ -79,6 +81,7 @@
 
 from skopt import BayesSearchCV
 from skopt.space import Real, Categorical, Integer
+from skopt.plots import plot_objective, plot_histogram
 
 from sklearn.datasets import load_digits
 from sklearn.svm import LinearSVC, SVC
@@ -114,14 +117,30 @@
 
 opt = BayesSearchCV(
     pipe,
-    [(svc_search, 20), (linsvc_search, 16)], # (parameter space, # of evaluations)
+    # (parameter space, # of evaluations)
+    [(svc_search, 40), (linsvc_search, 16)],
     cv=3
 )
 
 opt.fit(X_train, y_train)
 
 print("val. score: %s" % opt.best_score_)
 print("test score: %s" % opt.score(X_test, y_test))
+print("best params: %s" % str(opt.best_params_))
+
+#############################################################################
+# Partial Dependence plot
8000
 of the objective function for SVC
+#
+_ = plot_objective(opt.optimizer_results_[0],
+                   dimensions=["C", "degree", "gamma", "kernel"],
+                   n_minimum_search=int(1e8))
+plt.show()
+
+#############################################################################
+# Plot of the histogram for LinearSVC
+#
+_ = plot_histogram(opt.optimizer_results_[1], 1)
+plt.show()
 
 #############################################################################
 # Progress monitoring and control using `callback` argument of `fit` method
@@ -153,7 +172,6 @@
     cv=3
 )
 
-
 # callback handler
 def on_step(optim_result):
     score = searchcv.best_score_

diff --git a/skopt/plots.py b/skopt/plots.py
@@ -168,8 +168,10 @@ def plot_gaussian_process(res, **kwargs):
 
     if ax is None:
         ax = plt.gca()
-    assert res.space.n_dims == 1, "Space dimension must be 1"
-    x, x_model = _evenly_sample(res.space.dimensions[0], n_points)
+    n_dims = res.space.n_dims
+    assert n_dims == 1, "Space dimension must be 1"
+    dimension = res.space.dimensions[0]
+    x, x_model = _evenly_sample(dimension, n_points)
     x = x.reshape(-1, 1)
     x_model = x_model.reshape(-1, 1)
     if res.specs is not None and "args" in res.specs:
@@ -609,7 +611,8 @@ def plot_objective(result, levels=10, n_points=40, n_samples=250, size=2,
     plot_dims : list of str and int, default=None
         List of dimension names or dimension indices from the
         search-space dimensions to be included in the plot.
-        If `None` then use all dimensions from the search-space.
+        If `None` then use all dimensions except constant ones
+        from the search-space.
 
     sample_source : str or list of floats, default='random'
         Defines to samples generation to use for averaging the model function
@@ -623,31 +626,26 @@ def plot_objective(result, levels=10, n_points=40, n_samples=250, size=2,
 
         Valid strings:
 
-            - 'random' - `n_samples` random samples will used
-
-            - 'result' - Use only the best observed parameters
-
-            - 'expected_minimum' - Parameters that gives the best
-                  minimum Calculated using scipy's minimize method.
-                  This method currently does not work with categorical values.
-
-            - 'expected_minimum_random' - Parameters that gives the
-                  best minimum when using naive random sampling.
-                  Works with categorical values.
+        - 'random' - `n_samples` random samples will used
+        - 'result' - Use only the best observed parameters
+        - 'expected_minimum' - Parameters that gives the best
+          minimum Calculated using scipy's minimize method.
+          This method currently does not work with categorical values.
+        - 'expected_minimum_random' - Parameters that gives the
+          best minimum when using naive random sampling.
+          Works with categorical values.
 
     minimum : str or list of floats, default = 'result'
         Defines the values for the red points in the plots.
         Valid strings:
 
-            - 'result' - Use best observed parameters
-
-            - 'expected_minimum' - Parameters that gives the best
-                  minimum Calculated using scipy's minimize method.
-                  This method currently does not work with categorical values.
-
-            - 'expected_minimum_random' - Parameters that gives the
-                  best minimum when using naive random sampling.
-                  Works with categorical values
+        - 'result' - Use best observed parameters
+        - 'expected_minimum' - Parameters that gives the best
+          minimum Calculated using scipy's minimize method.
+          This method currently does not work with categorical values.
+        - 'expected_minimum_random' - Parameters that gives the
+          best minimum when using naive random sampling.
+          Works with categorical values
 
     n_minimum_search : int, default = None
         Determines how many points should be evaluated
@@ -660,6 +658,7 @@ def plot_objective(result, levels=10, n_points=40, n_samples=250, size=2,
     -------
     ax : `Matplotlib.Axes`
         A 2-d matrix of Axes-objects with the sub-plots.
+
     """
     # Here we define the values for which to plot the red dot (2d plot) and
     # the red dotted line (1d plot).
@@ -672,6 +671,8 @@ def plot_objective(result, levels=10, n_points=40, n_samples=250, size=2,
         # Get all dimensions.
         plot_dims = []
         for row in range(space.n_dims):
+            if space.dimensions[row].is_constant:
+                continue
             plot_dims.append((row, space.dimensions[row]))
     else:
         plot_dims = space[plot_dims]
@@ -771,12 +772,14 @@ def plot_evaluations(result, bins=20, dimensions=None,
     plot_dims : list of str and int, default=None
         List of dimension names or dimension indices from the
         search-space dimensions to be included in the plot.
-        If `None` then use all dimensions from the search-space.
+        If `None` then use all dimensions except constant ones
+        
F438
from the search-space.
 
     Returns
     -------
     ax : `Matplotlib.Axes`
         A 2-d matrix of Axes-objects with the sub-plots.
+
     """
     space = result.space
     # Convert categoricals to integers, so we can ensure consistent ordering.
@@ -790,6 +793,8 @@ def plot_evaluations(result, bins=20, dimensions=None,
         # Get all dimensions.
         plot_dims = []
         for row in range(space.n_dims):
+            if space.dimensions[row].is_constant:
+                continue
             plot_dims.append((row, space.dimensions[row]))
     else:
         plot_dims = space[plot_dims]
@@ -852,6 +857,7 @@ def _get_ylim_diagonal(ax):
     -------
     ylim_diagonal : tuple(int)
         The common min and max ylim for the diagonal plots.
+
     """
 
     # Number of search-space dimensions used in this plot.
@@ -913,6 +919,7 @@ def partial_dependence_1D(space, model, i, samples,
     yi : np.array
         The average value of the modelled objective function at
         each point `xi`.
+
     """
     # The idea is to step through one dimension, evaluating the model with
     # that dimension fixed and averaging either over random values or over

diff --git a/skopt/searchcv.py b/skopt/searchcv.py
@@ -549,6 +549,11 @@ def _make_optimizer(self, params_space):
         kwargs = self.optimizer_kwargs_.copy()
         kwargs['dimensions'] = dimensions_aslist(params_space)
         optimizer = Optimizer(**kwargs)
+        for i in range(len(optimizer.space.dimensions)):
+            if optimizer.space.dimensions[i].name is not None:
+                continue
+            optimizer.space.dimensions[i].name = list(sorted(
+                params_space.keys()))[i]
 
         return optimizer
 

diff --git a/skopt/space/space.py b/skopt/space/space.py
@@ -386,9 +386,9 @@ class Integer(Dimension):
         Distribution to use when sampling random integers for
         this dimension.
 
-        - If `"uniform"`, intgers are sampled uniformly between the lower
+        - If `"uniform"`, integers are sampled uniformly between the lower
           and upper bounds.
-        - If `"log-uniform"`, intgers are sampled uniformly between
+        - If `"log-uniform"`, integers are sampled uniformly between
           `log(lower, base)` and `log(upper, base)` where log
           has base `base`.
-Original file line number
+Diff line change
@@ Expand Up @@
             </div>
             <div class="overflow-hidden mx-2 text-center flex-fill">
               <a href="auto_examples/sklearn-gridsearchcv-replacement.html"  aria-label="sklearn-gridsearchcv-replacement">
+                <img src="_images/sphx_glr_sklearn-gridsearchcv-replacement_001.png" class="sk-index-img" alt="Scikit-learn hyperparameter search wrapper">
               </a>
             </div>
           </div>
@@ Expand Down @@