Merge branch 'sparse-kde' of https://github.com/GardevoirX/scikit-matter

 into sparse-kde

docs/src/references/neighbors.rst

@@ -9,3 +9,8 @@ Sparse Kernel Density Estimation
 --------------------------------
 
 .. autoclass:: skmatter.neighbors.SparseKDE
+    :show-inheritance:
+
+    .. automethod:: fit
+    .. automethod:: score_samples
+    .. automethod:: score

examples/neighbors/pamm.py

@@ -341,7 +341,7 @@ def cluster_distribution_3D(
 
 # %%
 # We need to estimate the probability at each grid point to do quick shift, which can
-# further partition the set of grid points in to several clusters. The resulting
+# further partition the set of grid points into several clusters. The resulting
 # clusters can be interpreted as (meta-)stable states of the system.
 #
 #
@@ -368,7 +368,7 @@ def cluster_distribution_3D(
 )
 
 # %%
-# Based on the results, the gaussian mixture model of the system can be generated:
+# Based on the results, the Gaussian mixture model of the system can be generated:
 #
 #
 
@@ -386,7 +386,7 @@ def cluster_distribution_3D(
 
 # %%
 # The final result shows seven (meta-)stable states of hydrogen bond. Here we also show
-# the reference hydrogen bond descriptor. The gaussian with the largest weight locates
+# the reference hydrogen bond descriptor. The Gaussian with the largest weight locates
 # closest to the reference point. This result shows that, with the help of the
 # `SparseKDE` and `QuickShift` algorithm, we can easily identify the (meta-)stable
 # states of the system objectively and without any prior knowledge about the system.
@@ -402,4 +402,4 @@ def cluster_distribution_3D(
 ax.scatter(REF_HB[0], REF_HB[1], REF_HB[2], marker="+", color="red", s=1000)
 
 # %%
-f"The gaussian with the highest probability is {np.argmax(cluster_weights) + 1}"
+f"The Gaussian with the highest probability is {np.argmax(cluster_weights) + 1}"

examples/neighbors/sparse-kde.py

@@ -8,7 +8,7 @@
 Example for the usage of the :class:`skmatter.neighbors.SparseKDE` class. This class is
 specifically designed for conducting pobabilistic analysis of molecular motifs
 (`PAMM <https://doi.org/10.1063/1.4900655>`_),
-which is quiet useful for analyzing motifs like H-bonds, coordination polyhedra, and
+which is quite useful for analyzing motifs like H-bonds, coordination polyhedra, and
 protein secondary structure.
 
 Here we show how to use the sparse KDE model to fit the probability distribution based
@@ -60,7 +60,7 @@
 plt.show()
 
 # %%
-# The perquisite of conducting sparse KDE is to partition the sample set. Here, we use
+# Sparse KDE requires a discretization of the sample space. Here, we use
 # the FPS method to generate grid points in the sample space:
 #
 #

src/skmatter/neighbors/_sparsekde.py

@@ -53,7 +53,7 @@ class SparseKDE(BaseEstimator):
         The fractional number of points in the voronoi cell of each grid points. Use
         this when each cell has a similar number of points.
     kernel : str, default=gaussian
-        The kernel used here. Now only the gaussian kernel is available.
+        The kernel used here. Now only the Gaussian kernel is available.
     verbose : bool, default=False
         Whether to print progress.
 
@@ -79,7 +79,7 @@ class SparseKDE(BaseEstimator):
     >>> np.random.seed(0)
     >>> n_samples = 10_000
 
-    To create two gaussians with different means and covariance and sample from them
+    To create two Gaussians with different means and covariance and sample from them
 
     >>> cov1 = [[1, 0.5], [0.5, 1]]
     >>> cov2 = [[1, 0.5], [0.5, 0.5]]