<s>
For	O
data	B-General_Concept
analysis	I-General_Concept
,	O
Random	B-General_Concept
mapping	I-General_Concept
(	O
RM	O
)	O
is	O
a	O
fast	O
dimensionality	B-Algorithm
reduction	I-Algorithm
method	O
categorized	O
as	O
feature	B-Algorithm
extraction	I-Algorithm
method	O
.	O
</s>
<s>
When	O
the	O
data	O
vectors	O
are	O
high-dimensional	O
it	O
is	O
computationally	O
infeasible	O
to	O
use	O
data	B-General_Concept
analysis	I-General_Concept
or	O
pattern	O
recognition	O
algorithms	O
which	O
repeatedly	O
compute	O
similarities	O
or	O
distances	O
in	O
the	O
original	O
data	O
space	O
.	O
</s>
<s>
In	O
a	O
text	B-Algorithm
mining	I-Algorithm
context	O
,	O
it	O
is	O
demonstrated	O
that	O
the	O
document	B-Algorithm
classification	I-Algorithm
accuracy	O
obtained	O
after	O
the	O
dimensionality	O
has	O
been	O
reduced	O
using	O
a	O
random	B-General_Concept
mapping	I-General_Concept
method	O
will	O
be	O
almost	O
as	O
good	O
as	O
the	O
original	O
accuracy	O
if	O
the	O
final	O
dimensionality	O
is	O
sufficiently	O
large	O
(	O
about	O
100	O
out	O
of	O
6000	O
)	O
.	O
</s>
