<s>
In	O
computational	O
linguistics	O
,	O
word-sense	B-General_Concept
induction	I-General_Concept
(	O
WSI	O
)	O
or	O
discrimination	O
is	O
an	O
open	O
problem	O
of	O
natural	B-Language
language	I-Language
processing	I-Language
,	O
which	O
concerns	O
the	O
automatic	O
identification	O
of	O
the	O
senses	O
of	O
a	O
word	O
(	O
i.e.	O
</s>
<s>
Given	O
that	O
the	O
output	O
of	O
word-sense	B-General_Concept
induction	I-General_Concept
is	O
a	O
set	O
of	O
senses	O
for	O
the	O
target	O
word	O
(	O
sense	O
inventory	O
)	O
,	O
this	O
task	O
is	O
strictly	O
related	O
to	O
that	O
of	O
word-sense	B-General_Concept
disambiguation	I-General_Concept
(	O
WSD	O
)	O
,	O
which	O
relies	O
on	O
a	O
predefined	O
sense	O
inventory	O
and	O
aims	O
to	O
solve	O
the	O
ambiguity	O
of	O
words	O
in	O
context	O
.	O
</s>
<s>
The	O
output	O
of	O
a	O
word-sense	B-General_Concept
induction	I-General_Concept
algorithm	O
is	O
a	O
clustering	B-Algorithm
of	O
contexts	O
in	O
which	O
the	O
target	O
word	O
occurs	O
or	O
a	O
clustering	B-Algorithm
of	O
words	O
related	O
to	O
the	O
target	O
word	O
.	O
</s>
<s>
A	O
well-known	O
approach	O
to	O
context	O
clustering	B-Algorithm
is	O
the	O
Context-group	O
Discrimination	O
algorithm	O
based	O
on	O
large	O
matrix	O
computation	O
methods	O
.	O
</s>
<s>
Word	O
clustering	B-Algorithm
is	O
a	O
different	O
approach	O
to	O
the	O
induction	O
of	O
word	O
senses	O
.	O
</s>
<s>
It	O
consists	O
of	O
clustering	B-Algorithm
words	O
,	O
which	O
are	O
semantically	O
similar	O
and	O
can	O
thus	O
bear	O
a	O
specific	O
meaning	O
.	O
</s>
<s>
Lin	O
’s	O
algorithm	O
is	O
a	O
prototypical	O
example	O
of	O
word	O
clustering	B-Algorithm
,	O
which	O
is	O
based	O
on	O
syntactic	O
dependency	O
statistics	O
,	O
which	O
occur	O
in	O
a	O
corpus	O
to	O
produce	O
sets	O
of	O
words	O
for	O
each	O
discovered	O
sense	O
of	O
a	O
target	O
word	O
.	O
</s>
<s>
The	O
Clustering	B-Algorithm
By	O
Committee	O
(	O
CBC	O
)	O
also	O
uses	O
syntactic	O
contexts	O
,	O
but	O
exploits	O
a	O
similarity	O
matrix	O
to	O
encode	O
the	O
similarities	O
between	O
words	O
and	O
relies	O
on	O
the	O
notion	O
of	O
committees	O
to	O
output	O
different	O
senses	O
of	O
the	O
word	O
of	O
interest	O
.	O
</s>
<s>
The	O
main	O
hypothesis	O
of	O
co-occurrence	O
graphs	O
assumes	O
that	O
the	O
semantics	O
of	O
a	O
word	O
can	O
be	O
represented	O
by	O
means	O
of	O
a	O
co-occurrence	O
graph	B-Application
,	O
whose	O
vertices	O
are	O
co-occurrences	O
and	O
edges	O
are	O
co-occurrence	O
relations	O
.	O
</s>
<s>
These	O
approaches	O
are	O
related	O
to	O
word	O
clustering	B-Algorithm
methods	O
,	O
where	O
co-occurrences	O
between	O
words	O
can	O
be	O
obtained	O
on	O
the	O
basis	O
of	O
grammatical	O
or	O
collocational	O
relations	O
.	O
</s>
<s>
HyperLex	O
is	O
the	O
successful	O
approaches	O
of	O
a	O
graph	B-Application
algorithm	O
,	O
based	O
on	O
the	O
identification	O
of	O
hubs	O
in	O
co-occurrence	O
graphs	O
,	O
which	O
have	O
to	O
cope	O
with	O
the	O
need	O
to	O
tune	O
a	O
large	O
number	O
of	O
parameters	O
.	O
</s>
<s>
To	O
deal	O
with	O
this	O
issue	O
several	O
graph-based	O
algorithms	O
have	O
been	O
proposed	O
,	O
which	O
are	O
based	O
on	O
simple	O
graph	B-Application
patterns	O
,	O
namely	O
Curvature	O
Clustering	B-Algorithm
,	O
Squares	O
,	O
Triangles	O
and	O
Diamonds	O
(	O
SquaT++	O
)	O
,	O
and	O
Balanced	O
Maximum	O
Spanning	O
Tree	O
Clustering	B-Algorithm
(	O
B-MST	O
)	O
.	O
</s>
<s>
The	O
patterns	O
aim	O
at	O
identifying	O
meanings	O
using	O
the	O
local	O
structural	O
properties	O
of	O
the	O
co-occurrence	O
graph	B-Application
.	O
</s>
<s>
A	O
randomized	O
algorithm	O
which	O
partitions	O
the	O
graph	B-Application
vertices	O
by	O
iteratively	O
transferring	O
the	O
mainstream	O
message	O
(	O
i.e.	O
</s>
<s>
Word-sense	B-General_Concept
induction	I-General_Concept
has	O
been	O
shown	O
to	O
benefit	O
Web	O
Information	O
Retrieval	O
when	O
highly	O
ambiguous	O
queries	O
are	O
employed	O
.	O
</s>
<s>
Simple	O
word-sense	B-General_Concept
induction	I-General_Concept
algorithms	O
boost	O
Web	B-Algorithm
search	I-Algorithm
result	I-Algorithm
clustering	I-Algorithm
considerably	O
and	O
improve	O
the	O
diversification	O
of	O
search	O
results	O
returned	O
by	O
search	O
engines	O
such	O
as	O
Yahoo	B-Application
!	I-Application
</s>
<s>
Word-sense	B-General_Concept
induction	I-General_Concept
has	O
been	O
applied	O
to	O
enrich	O
lexical	B-Application
resources	I-Application
such	O
as	O
WordNet	B-General_Concept
.	O
</s>
<s>
is	O
a	O
freely	O
available	O
open	O
source	O
software	O
package	O
that	O
performs	O
both	O
context	O
clustering	B-Algorithm
and	O
word	O
clustering	B-Algorithm
.	O
</s>
