<s>
In	O
computer	B-General_Concept
science	I-General_Concept
,	O
uncertain	B-Algorithm
data	I-Algorithm
is	O
data	O
that	O
contains	O
noise	O
that	O
makes	O
it	O
deviate	O
from	O
the	O
correct	O
,	O
intended	O
or	O
original	O
values	O
.	O
</s>
<s>
In	O
the	O
age	O
of	O
big	B-Application
data	I-Application
,	O
uncertainty	O
or	O
data	O
veracity	O
is	O
one	O
of	O
the	O
defining	O
characteristics	O
of	O
data	O
.	O
</s>
<s>
Uncertain	B-Algorithm
data	I-Algorithm
is	O
found	O
in	O
abundance	O
today	O
on	O
the	O
web	O
,	O
in	O
sensor	B-Architecture
networks	I-Architecture
,	O
within	O
enterprises	O
both	O
in	O
their	O
structured	O
and	O
unstructured	O
sources	O
.	O
</s>
<s>
In	O
2012	O
IBM	O
called	O
out	O
managing	O
uncertain	B-Algorithm
data	I-Algorithm
at	O
scale	O
in	O
its	O
global	O
technology	O
outlook	O
report	O
that	O
presents	O
a	O
comprehensive	O
analysis	O
looking	O
three	O
to	O
ten	O
years	O
into	O
the	O
future	O
seeking	O
to	O
identify	O
significant	O
,	O
disruptive	O
technologies	O
that	O
will	O
change	O
the	O
world	O
.	O
</s>
<s>
Analyses	O
based	O
on	O
uncertain	B-Algorithm
data	I-Algorithm
will	O
have	O
an	O
effect	O
on	O
the	O
quality	O
of	O
subsequent	O
decisions	O
,	O
so	O
the	O
degree	O
and	O
types	O
of	O
inaccuracies	O
in	O
this	O
uncertain	B-Algorithm
data	I-Algorithm
cannot	O
be	O
ignored	O
.	O
</s>
<s>
Uncertain	B-Algorithm
data	I-Algorithm
is	O
found	O
in	O
the	O
area	O
of	O
sensor	B-Architecture
networks	I-Architecture
;	O
text	O
where	O
noisy	O
text	O
is	O
found	O
in	O
abundance	O
on	O
social	O
media	O
,	O
web	O
and	O
within	O
enterprises	O
where	O
the	O
structured	O
and	O
unstructured	B-Application
data	I-Application
may	O
be	O
old	O
,	O
outdated	O
,	O
or	O
plain	O
incorrect	O
;	O
in	O
modeling	O
where	O
the	O
mathematical	O
model	O
may	O
only	O
be	O
an	O
approximation	O
of	O
the	O
actual	O
process	O
.	O
</s>
<s>
There	O
are	O
three	O
main	O
models	O
of	O
uncertain	B-Algorithm
data	I-Algorithm
in	O
databases	O
.	O
</s>
<s>
In	O
attribute	O
uncertainty	O
,	O
each	O
uncertain	O
attribute	O
in	O
a	O
tuple	B-Application
is	O
subject	O
to	O
its	O
own	O
independent	O
probability	O
distribution	O
.	O
</s>
<s>
In	O
tuple	B-Application
uncertainty	O
,	O
all	O
the	O
attributes	O
of	O
a	O
tuple	B-Application
are	O
subject	O
to	O
a	O
joint	O
probability	O
distribution	O
.	O
</s>
<s>
This	O
covers	O
the	O
case	O
of	O
correlated	O
uncertainty	O
,	O
but	O
also	O
includes	O
the	O
case	O
where	O
there	O
is	O
a	O
probability	O
of	O
a	O
tuple	B-Application
not	O
belonging	O
in	O
the	O
relevant	O
relation	O
,	O
which	O
is	O
indicated	O
by	O
all	O
the	O
probabilities	O
not	O
summing	O
to	O
one	O
.	O
</s>
<s>
For	O
example	O
,	O
assume	O
we	O
have	O
the	O
following	O
tuple	B-Application
from	O
a	O
probabilistic	B-General_Concept
database	I-General_Concept
:	O
</s>
<s>
Then	O
,	O
the	O
tuple	B-Application
has	O
10%	O
chance	O
of	O
not	O
existing	O
in	O
the	O
database	O
.	O
</s>
