General-purpose fit

you can almost fit any function you want

bisec_fit(data, ff, extra, space, partition, times, trim, enlarge)

Arguments

data	numeric matrix. data[i, 1] gives \(xi\) and data[i, 2] gives \(yi\)
ff	function. ff(para, extra) gives the function f() that \(y=f(x)\) where the range of para make a space.
extra	numeric vector. see ff
space, partition, times, trim, enlarge	see bisec_optim

Value

list of list.

1. information about subrange1

   1. \(R^2\) using median of subrange1 as para for ff

   2. subrange1

2. information about subrange2

   1. \(R^2\) using median of subrange2 as para for ff

   2. subrange1

3. ...

Details

even complex coupled differential equation that you can't get a analytic solution is ok.

to do

1. support multivariate function

See also

biosec_optim

Examples

bisec_fit(data1, f_protein, extra11, space1, 2, 5, 1/3, 1)
#> $value
#> [1] 0.0001447741
#> 
#> $para
#>       K.i     K.nat       k.d 
#> 0.8859375 0.0984375 0.0196875 
#> 
#> $space
#>          [,1]     [,2]
#> K.i   0.84375 0.928125
#> K.nat 0.09375 0.103125
#> k.d   0.01875 0.020625
#> 
bisec_fit(data2, f_AI2_out, extra2, space2, 2, 5, 4, 1)
#> $value
#> [1] 9.844207e-06
#> 
#> $para
#>       k.in      k.out        k.p 
#> 0.00787500 0.04429687 0.00646875 
#> 
#> $space
#>            [,1]       [,2]
#> k.in  0.0075000 0.00825000
#> k.out 0.0421875 0.04640625
#> k.p   0.0061875 0.00675000
#>