Fairness in constrained spectral clustering
Neurocomputing,
Год журнала:
2025,
Номер
634, С. 129815 - 129815
Опубликована: Март 1, 2025
Язык: Английский
One-Stage Fair Multi-View Spectral Clustering
Ruoyan Li,
Haiyang Hu,
Liang Du
и другие.
Опубликована: Окт. 26, 2024
Язык: Английский
Fair Clustering Ensemble With Equal Cluster Capacity
IEEE Transactions on Pattern Analysis and Machine Intelligence,
Год журнала:
2024,
Номер
47(3), С. 1729 - 1746
Опубликована: Ноя. 28, 2024
Clustering
ensemble
has
been
widely
studied
in
data
mining
and
machine
learning.
However,
the
existing
clustering
methods
do
not
pay
attention
to
fairness,
which
is
important
real-world
applications,
especially
applications
involving
humans.
To
address
this
issue,
paper
proposes
a
novel
fair
method,
takes
multiple
base
results
as
inputs
learns
consensus
result.
When
designing
algorithm,
we
observe
that
one
of
used
definitions
fairness
may
cause
cluster
imbalance
problem.
tackle
problem,
give
new
definition
can
simultaneously
characterize
capacity
equality.
Based
on
definition,
design
an
extremely
simple
yet
effective
regularized
term
achieve
We
plug
into
our
framework,
finally
leading
method.
The
extensive
experiments
show
that,
compared
with
state-of-the-art
methods,
method
only
comparable
or
even
better
performance,
but
also
obtain
much
fairer
equality
result,
well
demonstrates
effectiveness
superiority
Язык: Английский
Spatial Algorithmic Bias in Socio-Economic Clustering of Russian Regions
Spatial Economics,
Год журнала:
2024,
Номер
20(2), С. 71 - 92
Опубликована: Янв. 1, 2024
Decision-making
based
on
complex
human-machine
algorithms
can
lead
to
discrimination
of
citizens
gender,
race
and
other
grounds.
However,
in
world
science
there
is
no
idea
algorithmically
conditioned
by
their
place
residence.
This
also
applies
the
adoption
algorithmic
decisions
socio-economic
development
regions.
Therefore,
purpose
our
study
was
detect
bias
results
clustering
Russian
To
achieve
this
goal,
it
necessary
identify
sensitive
operations
cluster
analysis
that
could
spatial
injustice,
form
an
array
articles
subjects
(regions)
Federation,
analyze
all
for
possibility
regions
with
potentially
biased
attitudes
towards
them
as
a
result
clustering.
The
term
‘spatial
bias’
proposed.
Using
author’s
semantic
search
algorithm
bibliographic
databases,
six
hundred
empirical
indicators
were
identified.
characteristics
identified
are
given.
showed
most
evident
four
–
deploying
conceptual
model
into
optimal
set
indicators,
selecting
regions,
choosing
way
combine
clusters
determining
number
clusters.
Examples
discriminated
presented
each
operation.
Three
directions
further
research
indicated.
Practical
significance
may
be
associated
unbiased
regional
fair
Federation’s
Язык: Английский