fitMatLabCellML

Parameter estimation of CellML files in MatLab. Tool by David Cumin, ABI.

fitMatLabCellML is a MatLab utility that performs parameter estimation of CellML models. Note that the input format is the COR MatLab export format. To use this utility, load a CellML model in COR and export it as MatLab.

fitMatlabCellML1_5.m (together with DCsGUI.m, objectiveFunction.m, fminbndsearch.m (optional), and writeascii.m)
will read in a CellML model and allow a user to select which of the constant variables to use for the parameter estimation.

The objective function will be minimised. This must be written as a string of MatLab code, which returns a single value (e.g. if we wished to solve for x=5, then the objective function would read 'mean((5-x).^2)'). A temporary file will be created (the suffix of which can be specified,) and at the completion of the parameter estimation the new values will be displayed and
the user will be asked if you wish to save the new values in another m-file.

Variables:

time: A string for the operand. Some models use 't', some use 'time' etc
lengthToRun: The length of time you wish to run the model for (e.g. 0:0.1:10)
suffix: A string for the suffix of the temporary file (e.g. '_temp')
objF: The m-file name of the objective function (e.g. objectivefunction)
boundSearch: Boolean value: true if you want to bound the parameter estimation search (uses fminsearchbnd.m); false if you do not (uses fminsearch.m). fminsearch will converge slightly faster but may give unrealistic values.
constantMultipliers: if using fminsearchbnd.m, the multiple of each parameter to bound the search by (e.g. if this value is 2 then the search will be bounded to +-twice the value of each parameter).

 

Download fitMatLabCellML