From ftm
Jump to: navigation, search
(→‎Under Construction...)
(→‎Reference under construction...)
Line 28: Line 28:
 
   | descr=
 
   | descr=
 
   | arguments=0 - number of centers to use<br>
 
   | arguments=0 - number of centers to use<br>
−
   | attributes=ncenters - number of centers<br>criteria - criteria<br>mode - (diagonal|full|spherical) covariance computation types<br>outpriors - reference to external fmat to store priors<br>outcov - reference to external fmat to store covariance<br>outcenters - reference to external fmat to store centers<br>
+
   | attributes="ncenters - number of centers<br>criteria - criteria<br>mode - (diagonal|full|spherical) covariance computation types<br>outpriors - reference to external fmat to store priors<br>outcov - reference to external fmat to store covariance<br>outcenters - reference to external fmat to store centers<br>"
 
   | messages=outpriors - reference to external fmat to store priors<br>outcov - reference to external fmat to store covariance<br>outcenters - reference to external fmat to store centers<br>
 
   | messages=outpriors - reference to external fmat to store priors<br>outcov - reference to external fmat to store covariance<br>outcenters - reference to external fmat to store centers<br>
 
   | inlets=0 - fmat<br>
 
   | inlets=0 - fmat<br>

Revision as of 11:25, 15 November 2007

Reference under construction...

mnm.delta '
arguments: 0 - initialize the input size
1 - initialize the filter size
attributes: norm - normalization mode 1 (default) or 0
inadddel - add a delay to the delayed input
messages: clear - clear the memory of inputs
getdelay - get the input ring buffer
getring - get the input ring buffer
getnorm - get the normalization factor
getstate - get the internal weights vector
norm - normalization mode 1 (default) or 0
inadddel - add a delay to the delayed input
filtersize - set the filter size
insize - set the input size
inlets: 0 - input vector
outlets: 0 - output delayed inputs (in phase with deltas)
1 - output deltas
2 - internal values
mnm.diag '
arguments: none
attributes: out - reference to an external fmat to store the diagonal
messages: out - reference to an external fmat to store the diagonal
inlets: 0 - incoming fmat whose diagonal will be extracted
outlets: 0 - diagonal of the fmat
mnm.gmmem '
arguments: 0 - number of centers to use
attributes: "ncenters - number of centers
criteria - criteria
mode - (diagonal
messages: outpriors - reference to external fmat to store priors
outcov - reference to external fmat to store covariance
outcenters - reference to external fmat to store centers
inlets: 0 - fmat
outlets: 0 - fmat centers
1 - fmat covariance
2 - fmat priors
mnm.hist '
arguments: 0 - number of bins
attributes: norm - values: on=max, sum: normalise histogram so that max or sum is 1
bpf - output two-column fmat with bin indices and histogram values
messages: set_n - number of bins
norm - values: on=max, sum: normalise histogram so that max or sum is 1
bpf - output two-column fmat with bin indices and histogram values
inlets: 0 - data fmat or fvec
outlets: 0 - histogram fmat
1 - min data value
2 - max data value
mnm.lu '
arguments: none
attributes: outx - out X
outpivot - out pivot
outu - out U
outl - out L
messages: outx - out X
outpivot - out pivot
outu - out U
outl - out L
solve - solves system with incoming fmat and decomposed fmat
determinant - computes determinant of decomposed fmat
inlets: 0 - fmat to decompose
outlets: 0 - L
1 - U
2 - pivot
3 - X
4 - determinant
mnm.mahalanobis '
arguments: 0 - fmat of mean values
1 - fmat of covariance
attributes: out - reference to external fmat to store result
messages: out - reference to external fmat to store result
set_shift_weight - set shift weight per target (in) line
set_shift_weight - set shift weight per target (in) line
set_shift - set shift vector per database (mu) line
set_shift - set shift vector per database (mu) line
inlets: 0 - vector
1 - mu
2 - sigma
outlets: 0 - mahalanobis distance
mnm.meanstd '
arguments: 0 - (1
attributes: type - (fmat) always output an fmat even for scalar results
mode - (row
messages: outstd - reference to an external fmat to store the standard deviation
outmean - reference to an external fmat to store the mean
inlets: 0 - incoming fmat
outlets: 0 - mean
1 - standard deviation
mnm.minmax '
arguments: 0 - (1
attributes: type - (fmat) always output an fmat even for scalar results
mode - (row
messages: outargmax - reference to fmat to store the argmax
outmax - reference to fmat to store the max
outargmin - reference to fmat to store the argmin
outmin - reference to fmat to store the min
inlets: 0 - incoming fmat
outlets: 0 - min
1 - argmin
2 - max
3 - argmax
mnm.nmd '
arguments: none
attributes: itermax - itermax
sH - sH
criteria - criteria
outh - out H
messages: outh - out H
inlets: 0 - fmat
1 - fmat
2 -
outlets: 0 - fmat
mnm.nmf '
arguments: 0 - number of components
attributes: itermax - maximum number of iterations
rdim - number of components
criteria - (float) stopping criteria
outh - reference to external fmat to store H
outw - reference to external fmat to store W
messages: outh - reference to external fmat to store H
outw - reference to external fmat to store W
inlets: 0 - fmat to be decomposed
outlets: 0 - W
1 - H
mnm.qr '
arguments: none
attributes: outx - reference to external fmat to store X
outr - reference to external fmat to store R
outq - reference to external fmat to store Q
messages: outx - reference to external fmat to store X
outr - reference to external fmat to store R
outq - reference to external fmat to store Q
solve - solve system with QR decomposition
inlets: 0 - fmat to be decomposed
outlets: 0 - Q
1 - R
2 - X
mnm.stats '
arguments: output stats
attributes: norm - switch normalize
messages: set - set histogram vector (fmat)
clear - clear histogram
bang - output stats
norm - switch normalize
inlets: 0 - data input
outlets: 0 - average
1 - standard deviation
2 - count
mnm.submat '
arguments: none
attributes: out - reference to external fmat to store the sub-matrix
messages: out - reference to external fmat to store the sub-matrix
end - end of submatrix coordinates
begin - start of submatrix coordinates
inlets: 0 - fmat
outlets: 0 - sub-matrix
mnm.sum '
arguments: 0 - (1
attributes: type - (fmat) always output an fmat even for scalar results
mode - (row
messages: out - reference to external fmat to store the sum
inlets: 0 - incoming fmat
outlets: 0 - sum of fmat
mnm.svd '
arguments: 0 - number of singular values
attributes: mode - (auto
messages: outvt - reference to external fmat to store V'
outs - reference to external fmat to store S
outu - reference to external fmat to store U
inlets: 0 - fmat to be decomposed by SVD
outlets: 0 - U
1 - S
2 - V'
mnm.transpose '
arguments: none
attributes: out - reference to external fmat to store the transposition
messages: out - reference to external fmat to store the transposition
inlets: 0 - fmat to be transposed
outlets: 0 - transposed fmat
mnm.xdist2 '
arguments: 0 - fmat to be used as right operand
attributes: swap - (yes
messages: out - reference to external fmat to store the result
inlets: 0 - left operand
1 - right operand
outlets: 0 - resulting fmat
mnm.xmul '
arguments: 0 - fmat to be used as right operand
attributes: swap - (yes
messages: out - reference to an external fmat to store the result
inlets: 0 - left hand side fmat operand
1 - right hand side fmat operand
outlets: 0 - fmat result
Retrieved from "https://ftm2.ircam.fr/index.php?title=MnM_Modules&oldid=2278"