family_of_curves[source]
family_of_curves(curve_type,get_info, *varargin)
Provides curve specific computations like number of curve parameters, boundaries, log likelihood, etc. Only supports 'horz_indpnt' natively.
Arguments:
- curve_type: Family of curves, string, e.g. 'horz_indpnt'
- get_info: Cues for specific information / computation, string, e.g. 'get_nParams'
- varargin: Is either empty or has arguments depending on the computation
Returns the output of all computations.
horz_indpnt_curve[source]
horz_indpnt_curve(get_info,input_params)
Provides curve-specific computations for the curve type
horz_indpnt_curve.
Only called via family_of_curves within the toolbox.
Order of curve parameters: y1, x1, x2, y2, y3 and y4. x1 must always
precede x2 (when passing in curve parameters as well as when plotting).
compute_horz_indpnt_likelihood[source]
compute_horz_indpnt_likelihood(input_params)
Get the curve y-vals (with or without net effects) for each of the P particle curves and then compute the log probability mass function (pmf).
count_horz_indpnt_particles[source]
count_horz_indpnt_particles(input_params)
Use some criterion to carve out the curve space into theory consistent and theory inconsistent
get_horz_indpnt_curve_xy_vals[source]
get_horz_indpnt_curve_xy_vals(input_params)
This is the same as compute_likelihood in the sense map the predictor variable to the curve y val but there are some differences.
- We only know the curve parameters and we have to map all the x_values (0-to-1) to the curve y values where as in compute_likelihood we had specific x values (predictor variables) and curve parameters
- There is NO net effect cluster concept here 3. We DO NOT compute the pmf as well. Hence parts of the code will look similar but we felt these two chunks of code will need to be separate