<s>
In	O
statistical	O
hypothesis	O
testing	O
,	O
a	O
uniformly	B-General_Concept
most	I-General_Concept
powerful	I-General_Concept
(	O
UMP	O
)	O
test	O
is	O
a	O
hypothesis	O
test	O
which	O
has	O
the	O
greatest	O
power	B-General_Concept
among	O
all	O
possible	O
tests	O
of	O
a	O
given	O
size	O
α	O
.	O
</s>
<s>
For	O
example	O
,	O
according	O
to	O
the	O
Neyman	O
–	O
Pearson	O
lemma	O
,	O
the	O
likelihood-ratio	B-General_Concept
test	I-General_Concept
is	O
UMP	O
for	O
testing	O
simple	O
(	O
point	O
)	O
hypotheses	O
.	O
</s>
<s>
is	O
the	O
UMP	B-General_Concept
test	I-General_Concept
of	O
size	O
for	O
testing	O
vs	O
.	O
</s>
<s>
Finally	O
,	O
we	O
note	O
that	O
in	O
general	O
,	O
UMP	B-General_Concept
tests	I-General_Concept
do	O
not	O
exist	O
for	O
vector	O
parameters	O
or	O
for	O
two-sided	O
tests	O
(	O
a	O
test	O
in	O
which	O
one	O
hypothesis	O
lies	O
on	O
both	O
sides	O
of	O
the	O
alternative	O
)	O
.	O
</s>
<s>
As	O
a	O
result	O
,	O
no	O
test	O
is	O
uniformly	B-General_Concept
most	I-General_Concept
powerful	I-General_Concept
in	O
these	O
situations	O
.	O
</s>
