Merge pull request #1004 from gchq/feature/gkp_analytic_test

Analytic example and tests for Greedy Kernel Points

Author: gw265981

.cspell/library_terms.txt

@@ -72,6 +72,7 @@ logpdf
 lstsq
 mathbb
 mathbf
+mathcal
 mathrm
 matplotlib
 maxval

CHANGELOG.md

@@ -13,6 +13,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   (https://github.com/gchq/coreax/issues/934)
 - Added `reduce_iterative()` method to Kernel Herding. (https://github.com/gchq/coreax/pull/983)
 - Added probabilistic iterative Kernel Herding benchmarking results. (https://github.com/gchq/coreax/pull/983)
+- Analytic example with integration test for `GreedyKernelPoints` plus an analytic unit
+  test for the loss function. (https://github.com/gchq/coreax/pull/1004)
 
 ### Fixed
 
@@ -26,6 +28,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   (https://github.com/gchq/coreax/pull/1000)
 - Reduced the number of dimensions when applying UMAP in `pounce_benchmark.py`.
   (https://github.com/gchq/coreax/pull/1000)
+- Refactored `GreedyKernelPoints` and associated functions to make more extensible in
+  future. (https://github.com/gchq/coreax/pull/1004)
 
 ### Removed
 

coreax/kernels/scalar_valued.py

@@ -31,7 +31,7 @@ class LinearKernel(ScalarValuedKernel):
     r"""
     Define a linear kernel.
 
-    Given :math:`\rho`=` ``output_scale``and :math:`a=` ``constant`` the linear kernel
+    Given output scale :math:`\rho` and constant :math:`a`, the linear kernel
     is defined as :math:`k: \mathbb{R}^d\times \mathbb{R}^d \to \mathbb{R}`,
     :math:`k(x, y) = a + \rho (x)^T(y)`.
 

coreax/solvers/base.py

@@ -41,13 +41,13 @@ def reduce(
         self, dataset: _Data, solver_state: Optional[_State] = None
     ) -> tuple[_Coreset, _State]:
         r"""
-        Reduce 'dataset' to a coreset - solve the coreset problem.
+        Reduce ``dataset`` to a coreset - solve the coreset problem.
 
         :param dataset: The (potentially weighted and supervised) data to generate the
             coreset from
         :param solver_state: Solution state information, primarily used to cache
             expensive intermediate solution step information
-        :return: a tuple of the solved coreset and intermediate solver state information
+        :return: Solved coreset; Intermediate solver state information
         """