<s>
Multilinear	B-Algorithm
principal	I-Algorithm
component	I-Algorithm
analysis	I-Algorithm
(	O
MPCA	O
)	O
is	O
a	O
multilinear	O
extension	O
of	O
principal	B-Application
component	I-Application
analysis	I-Application
(	O
PCA	O
)	O
.	O
</s>
<s>
multilinear	O
tensor	O
models	O
,	O
such	O
as	O
multilinear	B-Algorithm
principal	I-Algorithm
component	I-Algorithm
analysis	I-Algorithm
(	O
MPCA	O
)	O
,	O
or	O
multilinear	O
independent	O
component	O
analysis	O
(	O
MICA	O
)	O
,	O
etc	O
.	O
</s>
<s>
The	O
origin	O
of	O
MPCA	O
can	O
be	O
traced	O
back	O
to	O
the	O
Tucker	B-Algorithm
decomposition	I-Algorithm
and	O
Peter	O
Kroonenberg	O
's	O
"	O
3-mode	O
PCA	O
"	O
work	O
.	O
</s>
<s>
In	O
2005	O
,	O
Vasilescu	O
and	O
Terzopoulos	O
introduced	O
the	O
Multilinear	B-Algorithm
PCA	I-Algorithm
terminology	O
as	O
a	O
way	O
to	O
better	O
differentiate	O
between	O
linear	O
and	O
multilinear	O
tensor	O
decomposition	O
,	O
as	O
well	O
as	O
,	O
to	O
better	O
differentiate	O
between	O
the	O
work	O
that	O
computed	O
2nd	O
order	O
statistics	O
associated	O
with	O
each	O
data	O
tensor	O
mode(axis )	O
,	O
and	O
subsequent	O
work	O
on	O
Multilinear	O
Independent	O
Component	O
Analysis	O
that	O
computed	O
higher	O
order	O
statistics	O
associated	O
with	O
each	O
tensor	O
mode/axis	O
.	O
</s>
<s>
Multilinear	B-Algorithm
PCA	I-Algorithm
may	O
be	O
applied	O
to	O
compute	O
the	O
causal	O
factors	O
of	O
data	O
formation	O
,	O
or	O
as	O
signal	O
processing	O
tool	O
on	O
data	O
tensors	O
whose	O
individual	O
observation	O
have	O
either	O
been	O
vectorized	O
,	O
<	O
ref	O
name	O
=	O
"	O
Vasilescu2003	O
"	O
>M.A.O.	O
</s>
