<s>
In	O
statistics	O
,	O
the	O
focused	B-General_Concept
information	I-General_Concept
criterion	I-General_Concept
(	O
FIC	O
)	O
is	O
a	O
method	O
for	O
selecting	O
the	O
most	O
appropriate	O
model	O
among	O
a	O
set	O
of	O
competitors	O
for	O
a	O
given	O
data	O
set	O
.	O
</s>
<s>
Unlike	O
most	O
other	O
model	O
selection	O
strategies	O
,	O
like	O
the	O
Akaike	O
information	O
criterion	O
(	O
AIC	O
)	O
,	O
the	O
Bayesian	B-General_Concept
information	I-General_Concept
criterion	I-General_Concept
(	O
BIC	B-General_Concept
)	O
and	O
the	O
deviance	O
information	O
criterion	O
(	O
DIC	O
)	O
,	O
the	O
FIC	O
does	O
not	O
attempt	O
to	O
assess	O
the	O
overall	O
fit	O
of	O
candidate	O
models	O
but	O
focuses	O
attention	O
directly	O
on	O
the	O
parameter	O
of	O
primary	O
interest	O
with	O
the	O
statistical	O
analysis	O
,	O
say	O
,	O
for	O
which	O
competing	O
models	O
lead	O
to	O
different	O
estimates	O
,	O
say	O
for	O
model	O
.	O
</s>
<s>
The	O
clearest	O
case	O
is	O
where	O
precision	O
is	O
taken	O
to	O
be	O
mean	B-Algorithm
squared	I-Algorithm
error	I-Algorithm
,	O
say	O
in	O
terms	O
of	O
squared	O
bias	O
and	O
variance	O
for	O
the	O
estimator	O
associated	O
with	O
model	O
.	O
</s>
<s>
FIC	O
formulae	O
are	O
then	O
available	O
in	O
a	O
variety	O
of	O
situations	O
,	O
both	O
for	O
handling	O
parametric	B-General_Concept
,	O
semiparametric	O
and	O
nonparametric	B-General_Concept
situations	O
,	O
involving	O
separate	O
estimation	O
of	O
squared	O
bias	O
and	O
variance	O
,	O
leading	O
to	O
estimated	O
precision	O
.	O
</s>
<s>
In	O
the	O
end	O
the	O
FIC	O
selects	O
the	O
model	O
with	O
smallest	O
estimated	O
mean	B-Algorithm
squared	I-Algorithm
error	I-Algorithm
.	O
</s>
