SpatialModule Class Reference

Module providing spatial methods for processes utilizing spatial information. More...

#include <spatial_module.hpp>

Inheritance diagram for SpatialModule:

Public Member Functions
	SpatialModule (const Data &data_, const Eigen::MatrixXi W_, double spatial_coeff, const Eigen::VectorXi *old_alloc_provider=nullptr, const std::unordered_map< int, std::vector< int > > *old_cluster_members_provider_=nullptr)
	Constructs a SpatialModule with parameter and data references.
Spatial Methods
Eigen::VectorXd	compute_similarity_obs (int obs_idx) const override
	Counts neighbors of an observation grouped by cluster membership.
double	compute_similarity_cls (int cls_idx, bool old_allo=false) const override
	Counts internal edges within a cluster.
double	compute_similarity_obs (int obs_idx, int cls_idx) const override
	Counts neighbors of an observation within a specific cluster.
Public Member Functions inherited from Module
	Module (const Eigen::VectorXi *old_allocations_provider_=nullptr, const std::unordered_map< int, std::vector< int > > *old_cluster_members_provider_=nullptr)
void	set_old_allocations_provider (const Eigen::VectorXi *provider)
void	set_old_cluster_members_provider (const std::unordered_map< int, std::vector< int > > *provider)
virtual	~Module ()=default

Protected Member Functions
void	neighbor_cache_compute ()
	Precomputes and stores neighbor indices for all observations.

Protected Attributes
std::vector< std::vector< int > >	neighbor_cache
	Cache storing neighbor indices for each observation.
const Data &	data_module
	Reference to data object with cluster assignments.
const Eigen::MatrixXi	W
	Reference to adjacency matrix W from covariates.
const double	spatial_weight = 1.0
	Weighting factor for spatial similarity.
Protected Attributes inherited from Module
const Eigen::VectorXi *	old_allocations_provider
	Provider function for accessing old allocation state.
const std::unordered_map< int, std::vector< int > > *	old_cluster_members_provider
	Provider function for accessing old cluster members map.

Detailed Description

Module providing spatial methods for processes utilizing spatial information.

This class offers utility functions to compute neighbor counts based on an adjacency matrix W, facilitating the incorporation of spatial dependencies in clustering algorithms. The neighbor relationships are cached at construction time for efficient repeated access.

Constructor & Destructor Documentation

SpatialModule()

SpatialModule::SpatialModule ( const Data & data_, const Eigen::MatrixXi W_, double spatial_coeff, const Eigen::VectorXi * old_alloc_provider = nullptr, const std::unordered_map< int, std::vector< int > > * old_cluster_members_provider_ = nullptr )

inline

Constructs a SpatialModule with parameter and data references.

Initializes the neighbor cache by calling neighbor_cache_compute().

Parameters

data_	Reference to the Data object with cluster assignments.
W_	Reference to the adjacency matrix W. matrix.
spatial_coeff	Weighting factor for spatial similarity.
old_alloc_provider	function to access old allocations for split-merge.
old_cluster_members_provider_	function to access old cluster members for split-merge.

Member Function Documentation

compute_similarity_cls()

double SpatialModule::compute_similarity_cls ( int cls_idx, bool old_allo = false ) const

overridevirtual

Counts internal edges within a cluster.

Computes the number of edges where both endpoints belong to the specified cluster. This is calculated as: (1/2) * sum_{i,j in cluster} W(i,j). The division by 2 accounts for the symmetry of W (each edge counted twice).

Parameters

cls_idx	The index of the cluster (0 to K-1).
old_allo	If true, uses old allocations from old_allocations_provider; if false, uses current allocations from data_module (default: false).

Returns

The total number of internal edges in the cluster.

Implements Module.

compute_similarity_obs() [1/2]

Eigen::VectorXd SpatialModule::compute_similarity_obs ( int obs_idx ) const

overridevirtual

Counts neighbors of an observation grouped by cluster membership.

Returns a vector where element k contains the number of neighbors of obs_idx that belong to cluster k.

Parameters

obs_idx

The index of the observation (0 to N-1).

Returns

A K-dimensional vector of neighbor counts per cluster.

Implements Module.

compute_similarity_obs() [2/2]

double SpatialModule::compute_similarity_obs ( int obs_idx, int cls_idx ) const

overridevirtual

Counts neighbors of an observation within a specific cluster.

Uses the cached neighbor list to efficiently count how many neighbors of observation obs_idx belong to cluster cls_idx.

Parameters

obs_idx	The index of the observation (0 to N-1).
cls_idx	The index of the cluster to consider for neighbor counting.

Returns

The number of neighbors for the observation in the specified cluster.

Implements Module.

neighbor_cache_compute()

void SpatialModule::neighbor_cache_compute ( )

protected

Precomputes and stores neighbor indices for all observations.

This method initializes neighbor_cache by extracting non-zero entries from each row of the adjacency matrix W. Called once during construction to enable O(neighbors) instead of O(N) lookup time.

Member Data Documentation

data_module

const Data& SpatialModule::data_module

protected

Reference to data object with cluster assignments.

neighbor_cache

std::vector<std::vector<int> > SpatialModule::neighbor_cache

protected

Cache storing neighbor indices for each observation.

neighbor_cache[i] contains a vector of indices j where W(i,j) == 1. This avoids repeatedly scanning the full adjacency matrix during MCMC sampling.

spatial_weight

const double SpatialModule::spatial_weight = 1.0

protected

Weighting factor for spatial similarity.

W

const Eigen::MatrixXi SpatialModule::W

protected

Reference to adjacency matrix W from covariates.

---

The documentation for this class was generated from the following files:

• src/processes/module/spatial_module.hpp
• src/processes/module/spatial_module.cpp