|
Transform
setup (Setup
button)
Setup has only one
independent control - block's name. Name should not
contain white spaces nor "\" symbols - these
characters may lead to mistakes in some cases.
Other controls are grouped in three tabs:
transformation setup, forward mapping and filter
setup and triggering setup.
Apply
button saves all parameters but doesn't start the
calculations;
Go!
button saves parameters and starts calculations, it
is enabled only in
Manual
trigger mode.
-
transformation setup
Linear Transformations
Note:
DataSet
connected to the input of the
Transform
block may be used as a
Data Source
(reference set used to calculate transformation
parameters)
simultaneously in all cases.
None: Doesn't perform any calculations.
Useful when only some mixing of the vector elements
is required. May be used as a buffer for
delaying data in loops.
N(0,1):
Normalization to zero-mean and unitary standard
deviation. Uses reference
DataSet
selected in the
Data Source
list to calculate mean values and standard
deviations for each position in the event input
vector. Obtained values are used to normalize
input vectors of the processed events:
outputi=(inputi
-
mi)/si.
SVD: Quotient
Singular Value Decomposition. Uses reference
DataSet
selected in the
Data Source
list to calculate mixing matrix that transforms
event's input vector from the original space to
the QSVD
space. This is an orthogonal
transformation. Axes of the QSVD space show the
directions where
e=ssignal/sbackground
have the largest values. In this implementation
axes are ordered in the descending order of
max(e,e-1).
Parameter
nComponents
selects how many most informative
dimensions you need in the resulting space
(note: there will be equal number of outputs
accessible in forwarding
tab). Transformation needs to know if the
reference event is a signal or a background -
this is recognized by the target value (target
vector must contain only one element in this
case): following values are considered as a
signal: 1.0, 0.95 and 0.9; and for a background
it is: 0.1, 0.05, 0.0, -0.9, -0.95, -1.0.
ICA:
Independent Component Analysis. Uses reference
DataSet
selected in the
Data Source
list to calculate mixing matrix that transforms
event's input vector from the original space to
the ICA
space. This is a non-orthogonal
transformation. Axes of the ICA space show the
directions where a measure of non-gausianity
have the largest values. Data distributions in
the directions of ICA axes are forced to be uncorrelated one to each
other, if this condition cannot
be met - transformation will be stopped and
message window will inform you about the
reason. No information about signal/background
is required and there are no restrictions on the
target vector length. Axes are ordered in
the descending order of kurtosis value
k=EX4 - 3(EX2)2
and you can select the number of the
most "independent" dimensions (nComponents
parameter).
TItem: Transformation based on the
reference
Transform
block selected in the
Data Source
list. Does exactly the same calculations as the
reference.
lock: Locks the
transformation's parameters. If checked when
QSVD or ICA is selected - mixing matrix is
calculated only once (if it was ready before
locking, the existing matrix will be preserved).
When project is saved to the file with locked
mixing matrix - this matrix is written to the
project file and it will be used next time the
file is opened. This is particularly useful for ICA decomposition, which may give different results on each run.
Fourier Transformation
FFTspace: Transformation of the event's
input vector to the frequency domain. Resulting
vector consists of an amplitude spectrum
followed by a phase spectrum.
FFTflt:
FFT filter, event's input vector is filtered
in the frequency domain and then inverse FFT is
performed. Frequencies below the
LoBound
and above
the
HiBound
threshold are removed.
clear input: Input
DataSets
will be cleared
after processing.
clear output:
Output
DataSets
will be cleared before processing. If
unchecked - new events will be attached to the
existing ones.
- forward
mapping and filter setup
When transformation is applied to an event, its
input vector is processed. Resulting output
vector is used only to compose vectors of the
destination events and it is not stored in the
source event's output vector. Destination events
in each output
DataSet
may be composed in a different way. Select a
DataSet
in the
Output Data Set
list and put desired
formatting
strings
for destination event vectors. In the image
above there is a simple example:
-
destination event input vector will be composed
of the first and second element of the source
event (temporary) output vector;
-
destination event output vector will be created
with the length=1 and it will be initialized to
0.0 (not-used entry "n"
without the index);
- destination event
target vector will be created from the first
element of the source event target vector;
- destination event not-used vector will not be
allocated.
Events stored in the output
DataSets
can be filtered (in the example from the image
above filter is turned off).
- triggering
setup
If
Manual
mode is selected,
Go!
button is enabled and it releases
calculations. If
Internal
mode is chosen -
Source
button is enabled and it allows to select source
of the trigger through the common
Connection Add
dialog window (Go!
button is disabled); processing starts when one
of the selected sources finished its own task.
--top--
--up-- |
