def compute_geom_brute(self, diffusion_time, n_neighbors):
data = self.data
dim = self.dim
#set radius according to paper (i think this is dim not d)
radius = (diffusion_time * (diffusion_time * np.pi * 4)**(dim/2))**(0.5)
#set adjacency radius large enough that points beyond it have affinity close to zero
bigradius = 3 * radius
adjacency_method = 'brute'
adjacency_kwds = {'radius':bigradius}
affinity_method = 'gaussian'
affinity_kwds = {'radius':radius}
laplacian_method = 'geometric'
laplacian_kwds = {'scaling_epps':1}
geom = Geometry(adjacency_method=adjacency_method, adjacency_kwds=adjacency_kwds,
affinity_method=affinity_method, affinity_kwds=affinity_kwds,
laplacian_method=laplacian_method, laplacian_kwds=laplacian_kwds)
geom.set_data_matrix(data)
geom.adjacency_matrix = geom.compute_adjacency_matrix()
geom.laplacian_matrix = geom.compute_affinity_matrix()
geom.laplacian_matrix = self.get_laplacian(geom, radius)
return(geom)
Maps which are also the most scalable method. A too small
radius can break some embedding methods specifically if there
are neighbors with no or too few neighbors.
You can check the connectedness of your affinity matrix
as follows:
'''
from scipy.sparse.csgraph import connected_components
rad1 = 0.2
# compute an adjacency matrix with a radius
geom.adjacency_kwds = {'radius':rad1}
adjacency_matrix = geom.compute_adjacency_matrix()
# compute the corresponding affinity matrix
geom.affinity_kwds = {'radius':rad1}
affinity_matrix = geom.compute_affinity_matrix({'radius':rad1})
(number_connected_components, labels) = connected_components(affinity_matrix)
print(number_connected_components)
'''
Since the number of connected components is large this indicates that
our radius is too small. Let's increase:
'''
rad2 = 0.5
# compute an adjacency matrix with a radius
geom.adjacency_kwds = {'radius':rad2}
adjacency_matrix = geom.compute_adjacency_matrix()
# compute the corresponding affinity matrix
geom.affinity_kwds = {'radius':rad2}
affinity_matrix = geom.compute_affinity_matrix({'radius':rad2})
(number_connected_components, labels) = connected_components(affinity_matrix)
Maps which are also the most scalable method. A too small
radius can break some embedding methods specifically if there
are neighbors with no or too few neighbors.
You can check the connectedness of your affinity matrix
as follows:
'''
from scipy.sparse.csgraph import connected_components
rad1 = 0.2
# compute an adjacency matrix with a radius
geom.adjacency_kwds = {'radius': rad1}
adjacency_matrix = geom.compute_adjacency_matrix()
# compute the corresponding affinity matrix
geom.affinity_kwds = {'radius': rad1}
affinity_matrix = geom.compute_affinity_matrix({'radius': rad1})
(number_connected_components, labels) = connected_components(affinity_matrix)
print(number_connected_components)
'''
Since the number of connected components is large this indicates that
our radius is too small. Let's increase:
'''
rad2 = 0.5
# compute an adjacency matrix with a radius
geom.adjacency_kwds = {'radius': rad2}
adjacency_matrix = geom.compute_adjacency_matrix()
# compute the corresponding affinity matrix
geom.affinity_kwds = {'radius': rad2}
affinity_matrix = geom.compute_affinity_matrix({'radius': rad2})
(number_connected_components, labels) = connected_components(affinity_matrix)
print(number_connected_components)