-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.txt
59 lines (51 loc) · 2.49 KB
/
README.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
Start AliasingScore app on MATLAB Command Line:
-----------------------------------------------
>> AliasingScore(model, tbl);
The input parameters are 'model', a SimBiology model, and 'tbl', a data
table used for fitting. Remark: the data table must contain a column
specifying groups, even if the data contains only one group.
Use AliasingScore app:
----------------------
The app has four panels:
* Data Table
------------
Display the data table 'tbl'.
* Data Info
------------
Edit column 'Data Association' to associate the data with a model
response, group variable and independent variable.
Edit column 'Defines Dose' to specify if a data column defines a dose.
Edit column 'Include in Aliasing Score' to specify which model response
should be included in the aliasing score; this usually includes all
measured responses.
* Component Info
----------------
Edit column 'Value' to change the values at which the aliasing score
is computed.
Edit column 'Include in Aliasing Score' to specify which model component
should be included in the aliasing score; this usually includes all
components to be fitted.
* Aliasing Score
----------------
The 'Simulate' button simulates the model and computes
the aliasing scores.
Use the 'Response' selection to switch between plots for
different responses.
Use the 'Show Aliasing Score For Current Response Only' to switch
between combined and response-wise aliasing scores.
Use the 'Stop Time' field to change the simulation stop time.
Use the 'Group' selection to switch between groups.
Use the 'Percentage' field to specify the relative variation of
parameters used to compute finite difference approximations of
sensitivities.
Remark: for accurate finite difference approximations, the
percentage should be set to a small value. If the normalized
response plots look asymmetric, then the finite difference
approximation is inaccurate.
Use the 'Use Sensitivity for Aliasing Score' to switch between
Simbiology's builtin sensitivity analysis and finite difference
approximation for computing the aliasing score.
Use the sliders to adjust the scan window used for the aliasing score.
Use the 'Zoom' checkbox to zoom in and out of plots.
Use the 'Use Data Times For Score' checkbox to only use measurement
times within the current scan window to compute the aliasing score.