<s>
Statistical	B-General_Concept
parsing	I-General_Concept
is	O
a	O
group	O
of	O
parsing	B-Language
methods	O
within	O
natural	B-Language
language	I-Language
processing	I-Language
.	O
</s>
<s>
Within	O
this	O
mindset	O
,	O
the	O
idea	O
of	O
associating	O
each	O
rule	O
with	O
a	O
probability	O
then	O
provides	O
the	O
relative	O
frequency	O
of	O
any	O
given	O
grammar	O
rule	O
and	O
,	O
by	O
deduction	O
,	O
the	O
probability	O
of	O
a	O
complete	O
parse	B-Language
for	O
a	O
sentence	O
.	O
</s>
<s>
(	O
The	O
probability	O
associated	O
with	O
a	O
grammar	O
rule	O
may	O
be	O
induced	O
,	O
but	O
the	O
application	O
of	O
that	O
grammar	O
rule	O
within	O
a	O
parse	B-Language
tree	O
and	O
the	O
computation	O
of	O
the	O
probability	O
of	O
the	O
parse	B-Language
tree	O
based	O
on	O
its	O
component	O
rules	O
is	O
a	O
form	O
of	O
deduction	O
.	O
)	O
</s>
<s>
Using	O
this	O
concept	O
,	O
statistical	O
parsers	B-Language
make	O
use	O
of	O
a	O
procedure	O
to	O
search	O
over	O
a	O
space	O
of	O
all	O
candidate	O
parses	B-Language
,	O
and	O
the	O
computation	O
of	O
each	O
candidate	O
's	O
probability	O
,	O
to	O
derive	O
the	O
most	O
probable	O
parse	B-Language
of	O
a	O
sentence	O
.	O
</s>
<s>
The	O
Viterbi	B-Algorithm
algorithm	I-Algorithm
is	O
one	O
popular	O
method	O
of	O
searching	O
for	O
the	O
most	O
probable	O
parse	B-Language
.	O
</s>
<s>
"	O
Search	O
"	O
in	O
this	O
context	O
is	O
an	O
application	O
of	O
search	B-Application
algorithms	I-Application
in	O
artificial	B-Application
intelligence	I-Application
.	I-Application
</s>
<s>
This	O
lack	O
of	O
meaning	O
is	O
not	O
seen	O
as	O
a	O
problem	O
by	O
most	O
linguists	O
(	O
for	O
a	O
discussion	O
on	O
this	O
point	O
,	O
see	O
Colorless	O
green	O
ideas	O
sleep	O
furiously	O
)	O
but	O
from	O
a	O
pragmatic	O
point	O
of	O
view	O
it	O
is	O
desirable	O
to	O
obtain	O
the	O
first	O
interpretation	O
rather	O
than	O
the	O
second	O
and	O
statistical	O
parsers	B-Language
achieve	O
this	O
by	O
ranking	O
the	O
interpretations	O
based	O
on	O
their	O
probability	O
.	O
</s>
<s>
There	O
are	O
a	O
number	O
of	O
methods	O
that	O
statistical	B-General_Concept
parsing	I-General_Concept
algorithms	O
frequently	O
use	O
.	O
</s>
<s>
Most	O
statistical	B-General_Concept
parsing	I-General_Concept
algorithms	O
are	O
based	O
on	O
a	O
modified	O
form	O
of	O
chart	B-Application
parsing	I-Application
.	O
</s>
<s>
The	O
modifications	O
are	O
necessary	O
to	O
support	O
an	O
extremely	O
large	O
number	O
of	O
grammatical	O
rules	O
and	O
therefore	O
search	O
space	O
,	O
and	O
essentially	O
involve	O
applying	O
classical	O
artificial	B-Application
intelligence	I-Application
algorithms	O
to	O
the	O
traditionally	O
exhaustive	O
search	O
.	O
</s>
<s>
Some	O
examples	O
of	O
the	O
optimisations	O
are	O
only	O
searching	O
a	O
likely	O
subset	O
of	O
the	O
search	O
space	O
(	O
stack	B-Algorithm
search	I-Algorithm
)	O
,	O
for	O
optimising	O
the	O
search	O
probability	O
(	O
Baum-Welch	O
algorithm	O
)	O
and	O
for	O
discarding	O
parses	B-Language
that	O
are	O
too	O
similar	O
to	O
be	O
treated	O
separately	O
(	O
Viterbi	B-Algorithm
algorithm	I-Algorithm
)	O
.	O
</s>
