python - fit multiple parametric curves with scipy -

python - fit multiple parametric curves with scipy -

i have set (at to the lowest degree 3) of curves (xy-data). each curve parameters e , t constant different. i'm searching coefficients a,n , m best fit on curves.

y= x/e + (a/n+1)*t^(n+1)*x^m

i tried curve_fit, have no thought how parameters e , t function f (see curve_fit documetation). furthermore i'm not sure if understand xdata correctly. doc says: m-length sequence or (k,m)-shaped array functions k predictors. what's predictor? ydata has 1 dimension can't feed multiple curves routine.

so curve_fit might wrong approach don't know magic words search right one. can't first 1 dealing problem.

any help appreciated. j.

one way utilize scipy.optimize.leastsq instead (curve_fit convenience wrapper around leastsq).

stack x info in 1 dimension; ditto y data. lengths of 3 individual datasets don't matter; let's phone call them n1, n2 , n3, new x , y have shape (n1+n2+n3,).

inside function optimize, can split info @ convenience. not nicest function, work:

def function(x, e, t, a, n, m): homecoming x/e + (a/n+1)*t^(n+1)*x^m def leastsq_function(params, *args): = params[0] n = params[1] m = params[2] x = args[0] y = args[1] e = args[2] t = args[3] n1, n2 = args[2] yfit = np.empty(x.shape) yfit[:n1] = function(x[:n1], e[0], t[0], a, n, m) yfit[n1:n2] = function(x[n1:n2], e[1], t[1], a, n, m) yfit[n2:] = function(x[n2:], e[2], t[2], a, n, m) homecoming y - yfit params0 = [a0, n0, m0] args = (x, y, (e0, e1, e2), (t0, t1, t2), (n1, n1+n2)) result = scipy.optimize.leastsq(leastsq_function, params0, args=args)

i have not tested this, principle. you're splitting info 3 different calls inside function optimized.

note scipy.optimize.leastsq requires function returns whatever value you'd minized, in case difference between actual y info , fitted function data. actual of import variables in leastsq parameters want fit for, not x , y data. latter passed arguments, sizes of 3 separate datasets (i'm not using n3, , i've done juggling n1+n2 convenience; maintain in mind n1 , n2 within leastsq_function local variables, not original ones).

since awkward function fit (it won't have smooth derivative, example), quite essential

provide starting values (params0, ...0 values).

don't have info or parameters span orders of magnitude. closer around 1 (a few orders of magnitude ok), better.

python curve-fitting