arXiv (Cornell University),
Journal Year:
2023,
Volume and Issue:
unknown
Published: Jan. 1, 2023
Many
networked
datasets
with
units
interacting
in
groups
of
two
or
more,
encoded
hypergraphs,
are
accompanied
by
extra
information
about
nodes,
such
as
the
role
an
individual
a
workplace.
Here
we
show
how
these
node
attributes
can
be
used
to
improve
our
understanding
structure
resulting
from
higher-order
interactions.
We
consider
problem
community
detection
hypergraphs
and
develop
principled
model
that
combines
interactions
better
represent
observed
detect
communities
more
accurately
than
using
either
types
alone.
The
method
learns
automatically
input
data
extent
which
contribute
explain
data,
down
weighing
discarding
if
not
informative.
Our
algorithmic
implementation
is
efficient
scales
large
numbers
units.
apply
variety
systems,
showing
strong
performance
hyperedge
prediction
tasks
selecting
divisions
correlate
when
informative,
but
them
otherwise.
approach
illustrates
advantage
informative
available
data.
Nature Communications,
Journal Year:
2024,
Volume and Issue:
15(1)
Published: Aug. 16, 2024
Abstract
Many
networked
datasets
with
units
interacting
in
groups
of
two
or
more,
encoded
hypergraphs,
are
accompanied
by
extra
information
about
nodes,
such
as
the
role
an
individual
a
workplace.
Here
we
show
how
these
node
attributes
can
be
used
to
improve
our
understanding
structure
resulting
from
higher-order
interactions.
We
consider
problem
community
detection
hypergraphs
and
develop
principled
model
that
combines
interactions
better
represent
observed
detect
communities
more
accurately
than
using
either
types
alone.
The
method
learns
automatically
input
data
extent
which
contribute
explain
data,
down
weighing
discarding
if
not
informative.
Our
algorithmic
implementation
is
efficient
scales
large
numbers
units.
apply
variety
systems,
showing
strong
performance
hyperedge
prediction
tasks
selecting
divisions
correlate
when
informative,
but
them
otherwise.
approach
illustrates
advantage
informative
available
data.
Information,
Journal Year:
2025,
Volume and Issue:
16(4), P. 324 - 324
Published: April 18, 2025
Historical
figures
are
crucial
for
understanding
historical
processes
and
social
changes.
However,
existing
databases
of
primarily
focused
on
ancient
Chinese
individuals
limited
by
the
simplistic
organization
textual
information,
lacking
structured
processing.
Therefore,
this
study
proposes
an
automatic
method
constructing
a
spatio-temporal
database
modern
figures.
The
character
state
transition
matrix
reveals
evolution
figures,
while
random
walk
algorithm
identifies
their
primary
migration
patterns.
Using
from
Fujian
Province
(1840–2009)
as
case
study,
results
demonstrate
that
effectively
constructs
chain
encompassing
time,
space,
events.
indicates
fluctuating
trend
change
1840
to
2009,
initially
increasing
then
decreasing.
By
applying
keyword
extraction
method,
finds
transitions
causes
align
with
trends.
four-dimensional
analytical
framework
“character-time-space-event”
established
in
holds
significant
value
field
digital
humanities.
Scientific Reports,
Journal Year:
2024,
Volume and Issue:
14(1)
Published: March 23, 2024
The
hypergraph
community
detection
problem
seeks
to
identify
groups
of
related
vertices
in
data.
We
propose
an
information-theoretic
algorithm
which
compresses
the
observed
data
terms
labels
and
community-edge
intersections.
This
can
also
be
viewed
as
maximum-likelihood
inference
a
degree-corrected
microcanonical
stochastic
blockmodel.
perform
compression/inference
step
via
simulated
annealing.
Unlike
several
recent
algorithms
based
on
canonical
models,
our
does
not
require
statistical
parameters
such
vertex
degrees
or
pairwise
group
connection
rates.
Through
synthetic
experiments,
we
find
that
succeeds
down
recently-conjectured
thresholds
for
sparse
random
hypergraphs.
competitive
performance
cluster
recovery
tasks
sets.
Journal of Statistical Mechanics Theory and Experiment,
Journal Year:
2024,
Volume and Issue:
2024(4), P. 043403 - 043403
Published: April 23, 2024
Abstract
Hypergraphs
are
widely
adopted
tools
to
examine
systems
with
higher-order
interactions.
Despite
recent
advancements
in
methods
for
community
detection
these
systems,
we
still
lack
a
theoretical
analysis
of
their
detectability
limits.
Here,
derive
closed-form
bounds
hypergraphs.
Using
message-passing
formulation,
demonstrate
that
depends
on
the
hypergraphs’
structural
properties,
such
as
distribution
hyperedge
sizes
or
assortativity.
Our
formulation
enables
characterization
entropy
hypergraph
relation
its
clique
expansion,
showing
is
enhanced
when
hyperedges
highly
overlap
pairs
nodes.
We
develop
an
efficient
algorithm
learn
communities
and
model
parameters
large
systems.
Additionally,
devise
exact
sampling
routine
generate
synthetic
data
from
our
probabilistic
model.
methods,
numerically
investigate
boundaries
datasets,
extract
real
results
extend
understanding
limits
hypergraphs
introduce
flexible
mathematical
study
Applied Network Science,
Journal Year:
2024,
Volume and Issue:
9(1)
Published: Sept. 3, 2024
Abstract
A
wide
variety
of
complex
systems
are
characterized
by
interactions
different
types
involving
varying
numbers
units.
Multiplex
hypergraphs
serve
as
a
tool
to
describe
such
structures,
capturing
distinct
higher-order
among
collection
In
this
work,
we
introduce
comprehensive
set
measures
structural
connectivity
patterns
in
multiplex
hypergraphs,
considering
scales
from
node
and
hyperedge
levels
the
system’s
mesoscale.
We
validate
our
with
three
real-world
datasets:
scientific
co-authorship
physics,
movie
collaborations,
high
school
interactions.
This
validation
reveals
new
collaboration
patterns,
identifies
trends
within
across
subfields,
provides
insights
into
daily
interaction
dynamics.
Our
framework
aims
offer
more
nuanced
characterization
marked
both
