<s>
In	O
mathematics	O
,	O
a	O
Relevance	B-General_Concept
Vector	I-General_Concept
Machine	I-General_Concept
(	O
RVM	O
)	O
is	O
a	O
machine	O
learning	O
technique	O
that	O
uses	O
Bayesian	O
inference	O
to	O
obtain	O
parsimonious	O
solutions	O
for	O
regression	O
and	O
probabilistic	B-General_Concept
classification	I-General_Concept
.	O
</s>
<s>
The	O
RVM	O
has	O
an	O
identical	O
functional	O
form	O
to	O
the	O
support	B-Algorithm
vector	I-Algorithm
machine	I-Algorithm
,	O
but	O
provides	O
probabilistic	B-General_Concept
classification	I-General_Concept
.	O
</s>
<s>
It	O
is	O
actually	O
equivalent	O
to	O
a	O
Gaussian	B-General_Concept
process	I-General_Concept
model	O
with	O
covariance	O
function	O
:	O
</s>
<s>
Compared	O
to	O
that	O
of	O
support	B-Algorithm
vector	I-Algorithm
machines	I-Algorithm
(	O
SVM	B-Algorithm
)	O
,	O
the	O
Bayesian	O
formulation	O
of	O
the	O
RVM	O
avoids	O
the	O
set	O
of	O
free	O
parameters	O
of	O
the	O
SVM	B-Algorithm
(	O
that	O
usually	O
require	O
cross-validation-based	O
post-optimizations	O
)	O
.	O
</s>
<s>
However	O
RVMs	O
use	O
an	O
expectation	B-Algorithm
maximization	I-Algorithm
(	O
EM	O
)	O
-like	O
learning	O
method	O
and	O
are	O
therefore	O
at	O
risk	O
of	O
local	O
minima	O
.	O
</s>
<s>
This	O
is	O
unlike	O
the	O
standard	O
sequential	B-Algorithm
minimal	I-Algorithm
optimization	I-Algorithm
(	O
SMO	O
)	O
-based	O
algorithms	O
employed	O
by	O
SVMs	B-Algorithm
,	O
which	O
are	O
guaranteed	O
to	O
find	O
a	O
global	O
optimum	O
(	O
of	O
the	O
convex	O
problem	O
)	O
.	O
</s>
<s>
The	O
relevance	B-General_Concept
vector	I-General_Concept
machine	I-General_Concept
was	O
patented	O
in	O
the	O
United	O
States	O
by	O
Microsoft	O
(	O
patent	O
expired	O
September	O
4	O
,	O
2019	O
)	O
.	O
</s>
