初期値をきめるために。
>> help PAL_PFML_Fit
PAL_PFML_Fit Fit a psychometric function to data using a Maximum
Likelihood criterion.
:
'searchGrid': Either a 1x4 vector containing initial guesses for free
parametervalues and fixed values for fixed parameters [threshold
slope guess-rate lapse-rate] or a structure with vector fields
.alpha, .beta, .gamma, .lambda collectively defining a 4D parameter
grid through which to perform a brute-force search for initial
guesses (using PAL_PFML_BruteForceFit). Fields for fixed parameters
should be scalars equal to the fixed value. Note that choices made
here have a large effect on processing time and memory usage.
'paramsFree': 1x4 vector coding which of the four parameters in
paramsValue are free parameters and which are fixed parameters
(1: free, 0: fixed).
:
Full example:
options = optimset('fminsearch');
PF = @PAL_Logistic;
StimLevels = [-3:1:3];
NumPos = [55 55 66 75 91 94 97]; %observer data
OutOfNum = 100.*ones(size(StimLevels));
searchGrid.alpha = [-1:.01:1]; %structure defining grid to <--- この部分以下で範囲を設定
searchGrid.beta = 10.^[-1:.01:2]; %search for initial values
searchGrid.gamma = .5;
searchGrid.lambda = [0:.005:.06];
%or (not advised):
% searchGrid = [0 1 .5 0.01]; %Guesses
paramsFree = [1 1 0 1]
: