groupregr_f
Finds parameter values for a group of pets that minimizes the lossfunction using Nelder Mead's simplex method using a filter
Contents
Syntax
[q, info, itercount] = groupregr_f (func, par, data, auxData, weights, weightsPar, filternm)
Description
Finds parameter values for a group of pets that minimizes the lossfunction using Nelder Mead's simplex method using a filter. The filter gives always a pass in the case that no filter has been selected in estim_options. Parameters can be shared (scalar), or free (vector) with weighted reduction of scaled variances.
Input
- func: character string with name of user-defined function that specifies predictions, given parameters; see nrregr_st or nrregr
- par: structure with parameters in the format as specified in pars_init_group
- data: structure with data
- auxData: structure with auxiliary data
- weights: structure with weights for data
- weightsPar: structure with weights for parameters to minimize scaled variances
- filternm: character or cell string with name of user-defined filter function
Output
- q: structure with parameters, result of minimization of the lossfunction
- info: 1 if convergence has been successful; 0 otherwise
- itercount: number if iterations
Remarks
Set options with nmregr_options. Similar to nrregr_st, but slower and a larger bassin of attraction and uses a filter. The number of fields in data is variable. See petregr_f for the single-species situation