Modeling:Amps

From QuB

Jump to: navigation, search
Prev: Modeling:Idealize Outline Next: Modeling:Idl/Base


Amps (AMP) re-estimates the amplitude and standard deviation of all conductance classes in the Model, using the Baum-Welch algorithm. The new amp and std values are put into the model.

AMP can produce an idealization of the data, by picking statet that maximizes Γstate,t = Forwardstate,t * Backwardstate,t. However, the Modeling:Idealize button generally idealizes as well, but much faster, and with unlimited-length data. AMP should be used on small selections of data, and its idealization should just be a quick visual check that it worked right.

AMP also produces all the same measurements and histograms as Modeling:Idealize.

You can run AMP directly from the Model window: select some Data, right-click in the background of the Model window, and choose "Grab all amps increasing/decreasing." This has the same effect as Data Source:sel, Auto-Init:up/down, don't Show Idealized, Quiet (no text) Output. No Results are shown.

Properties

Data channel index which A/D channel contains the data to measure, typically 0
Data source
Pre-process data none: use raw data
as displayed: use the same baseline correction and filtering as the Data display
as such: apply the specified filter and baseline
Hist bin count number of bins for amplitude histograms, displayed in the Results window
Auto-Init initialize mean amplitudes by spacing them evenly between min(data) and max(data)
Open is up/down when auto-initializing, should the open current be greater than (up) or less than (down) the closed current?
Show Idealized generate an idealization in the Data window
Quiet Output don't print anything in the Report window

Results

In the textual Report window: the re-estimated amplitudes and standard deviations.

In the Results window:

Summary:

amp the mean amplitude of each class
sd the standard deviation of each class

Segments (and Select, Criteria): all the measurements of each segment/selection

Models: the model used

Histograms: An amplitude histogram for each segment/selection, overlaid with the Gaussian distribution (amp, std) of each class. You can quickly identify segments with errors by looking for histograms whose distribution curves don't match the histogram bars.


Prev: Modeling:Idealize Outline Next: Modeling:Idl/Base