<s>
Youden	B-General_Concept
's	I-General_Concept
J	I-General_Concept
statistic	I-General_Concept
(	O
also	O
called	O
Youden	B-General_Concept
's	I-General_Concept
index	I-General_Concept
)	O
is	O
a	O
single	O
statistic	O
that	O
captures	O
the	O
performance	O
of	O
a	O
dichotomous	O
diagnostic	O
test	O
.	O
</s>
<s>
Informedness	B-General_Concept
is	O
its	O
generalization	O
to	O
the	O
multiclass	O
case	O
and	O
estimates	O
the	O
probability	O
of	O
an	O
informed	O
decision	O
.	O
</s>
<s>
Youden	B-General_Concept
's	I-General_Concept
index	I-General_Concept
is	O
often	O
used	O
in	O
conjunction	O
with	O
receiver	B-Algorithm
operating	I-Algorithm
characteristic	I-Algorithm
(	O
ROC	O
)	O
analysis	O
.	O
</s>
<s>
The	O
index	O
is	O
defined	O
for	O
all	O
points	O
of	O
an	O
ROC	B-Algorithm
curve	I-Algorithm
,	O
and	O
the	O
maximum	O
value	O
of	O
the	O
index	O
may	O
be	O
used	O
as	O
a	O
criterion	O
for	O
selecting	O
the	O
optimum	O
cut-off	O
point	O
when	O
a	O
diagnostic	O
test	O
gives	O
a	O
numeric	O
rather	O
than	O
a	O
dichotomous	O
result	O
.	O
</s>
<s>
Youden	B-General_Concept
's	I-General_Concept
index	I-General_Concept
is	O
also	O
known	O
as	O
deltaP	O
 '	O
and	O
generalizes	O
from	O
the	O
dichotomous	O
to	O
the	O
multiclass	O
case	O
as	O
informedness	B-General_Concept
.	O
</s>
<s>
The	O
use	O
of	O
a	O
single	O
index	O
is	O
"	O
not	O
generally	O
to	O
be	O
recommended	O
"	O
,	O
but	O
informedness	B-General_Concept
or	O
Youden	B-General_Concept
's	I-General_Concept
index	I-General_Concept
is	O
the	O
probability	O
of	O
an	O
informed	O
decision	O
(	O
as	O
opposed	O
to	O
a	O
random	O
guess	O
)	O
and	O
takes	O
into	O
account	O
all	O
predictions	O
.	O
</s>
<s>
An	O
unrelated	O
but	O
commonly	O
used	O
combination	O
of	O
basic	O
statistics	O
from	O
information	B-Library
retrieval	I-Library
is	O
the	O
F-score	B-General_Concept
,	O
being	O
a	O
(	O
possibly	O
weighted	O
)	O
harmonic	O
mean	O
of	O
recall	O
and	O
precision	O
where	O
recall	O
=	O
sensitivity	O
=	O
true	O
positive	O
rate	O
,	O
but	O
specificity	O
and	O
precision	O
are	O
totally	O
different	O
measures	O
.	O
</s>
<s>
F-score	B-General_Concept
,	O
like	O
recall	O
and	O
precision	O
,	O
only	O
considers	O
the	O
so-called	O
positive	O
predictions	O
,	O
with	O
recall	O
being	O
the	O
probability	O
of	O
predicting	O
just	O
the	O
positive	O
class	O
,	O
precision	O
being	O
the	O
probability	O
of	O
a	O
positive	O
prediction	O
being	O
correct	O
,	O
and	O
F-score	B-General_Concept
equating	O
these	O
probabilities	O
under	O
the	O
effective	O
assumption	O
that	O
the	O
positive	O
labels	O
and	O
the	O
positive	O
predictions	O
should	O
have	O
the	O
same	O
distribution	O
and	O
prevalence	O
,	O
similar	O
to	O
the	O
assumption	O
underlying	O
of	O
Fleiss	O
 '	O
kappa	O
.	O
</s>
<s>
Youden	B-General_Concept
's	I-General_Concept
J	I-General_Concept
,	O
Informedness	B-General_Concept
,	O
Recall	O
,	O
Precision	O
and	O
F-score	B-General_Concept
are	O
intrinsically	O
undirectional	O
,	O
aiming	O
to	O
assess	O
the	O
deductive	O
effectiveness	O
of	O
predictions	O
in	O
the	O
direction	O
proposed	O
by	O
a	O
rule	O
,	O
theory	O
or	O
classifier	O
.	O
</s>
<s>
Markedness	O
(	O
deltaP	O
)	O
is	O
Youden	B-General_Concept
's	I-General_Concept
J	I-General_Concept
used	O
to	O
assess	O
the	O
reverse	O
or	O
abductive	O
direction	O
,	O
and	O
matches	O
well	O
human	O
learning	O
of	O
associations	O
;	O
rules	O
and	O
,	O
superstitions	O
as	O
we	O
model	O
possible	O
causation	B-Application
;	O
while	O
correlation	O
and	O
kappa	O
evaluate	O
bidirectionally	O
.	O
</s>
<s>
Matthews	B-General_Concept
correlation	I-General_Concept
coefficient	I-General_Concept
is	O
the	O
geometric	O
mean	O
of	O
the	O
regression	B-General_Concept
coefficient	I-General_Concept
of	O
the	O
problem	O
and	O
its	O
dual	O
,	O
where	O
the	O
component	O
regression	B-General_Concept
coefficients	I-General_Concept
of	O
the	O
Matthews	B-General_Concept
correlation	I-General_Concept
coefficient	I-General_Concept
are	O
Markedness	O
(	O
inverse	O
of	O
Youden	B-General_Concept
's	I-General_Concept
J	I-General_Concept
or	O
deltaP	O
)	O
and	O
informedness	B-General_Concept
(	O
Youden	B-General_Concept
's	I-General_Concept
J	I-General_Concept
or	O
deltaP	O
 '	O
)	O
.	O
</s>
<s>
Kappa	B-General_Concept
statistics	I-General_Concept
such	O
as	O
Fleiss	O
 '	O
kappa	O
and	O
Cohen	B-General_Concept
's	I-General_Concept
kappa	I-General_Concept
are	O
methods	O
for	O
calculating	O
inter-rater	O
reliability	O
based	O
on	O
different	O
assumptions	O
about	O
the	O
marginal	O
or	O
prior	O
distributions	O
,	O
and	O
are	O
increasingly	O
used	O
as	O
chance	O
corrected	O
alternatives	O
to	O
accuracy	O
in	O
other	O
contexts	O
.	O
</s>
<s>
Fleiss	O
 '	O
kappa	O
,	O
like	O
F-score	B-General_Concept
,	O
assumes	O
that	O
both	O
variables	O
are	O
drawn	O
from	O
the	O
same	O
distribution	O
and	O
thus	O
have	O
the	O
same	O
expected	O
prevalence	O
,	O
while	O
Cohen	B-General_Concept
's	I-General_Concept
kappa	I-General_Concept
assumes	O
that	O
the	O
variables	O
are	O
drawn	O
from	O
distinct	O
distributions	O
and	O
referenced	O
to	O
a	O
model	O
of	O
expectation	O
that	O
assumes	O
prevalences	O
are	O
independent	O
.	O
</s>
<s>
When	O
the	O
true	O
prevalences	O
for	O
the	O
two	O
positive	O
variables	O
are	O
equal	O
as	O
assumed	O
in	O
Fleiss	O
kappa	O
and	O
F-score	B-General_Concept
,	O
that	O
is	O
the	O
number	O
of	O
positive	O
predictions	O
matches	O
the	O
number	O
of	O
positive	O
classes	O
in	O
the	O
dichotomous	O
(	O
two	O
class	O
)	O
case	O
,	O
the	O
different	O
kappa	O
and	O
correlation	O
measure	O
collapse	O
to	O
identity	O
with	O
Youden	B-General_Concept
's	I-General_Concept
J	I-General_Concept
,	O
and	O
recall	O
,	O
precision	O
and	O
F-score	B-General_Concept
are	O
similarly	O
identical	O
with	O
accuracy	O
.	O
</s>
