python 3.x - new error in old code in numpy exp -

recently working on data able obtain curve using curve_fit after saving plot , values obtained returned same code later find not work.

#! python 3.5.2  import numpy np import matplotlib.pyplot plt import scipy.stats scipy.optimize import curve_fit  data= np.array([     [24, 0.176644513],     [27, 0.146382841],     [30, 0.129891534],     [33, 0.105370908],     [38, 0.077820511],     [50, 0.047407538]]) x, y = np.array([]), np.array([])  val in data:     x = np.append(x, val[0])     y = np.append(y, (val[1]/(1-val[1])))  def f(x, a, b):      return (np.exp(-a*x)**b)  # original , b values obtained = -0.2 # after rounding b = -0.32 # after rounding plt.scatter(x, y) xcurve = np.linspace(x[0], x[-1], 500) plt.plot(xcurve, f(xcurve,a,b), ls='--', color='k', lw=1) plt.show()  # original code values = b = 1 popt, pcov = curve_fit(f, x, y, (a, b)) 

whereas, curve_fit returned values a, b = -0.2, -0.32 returns: warning (from warnings module): file "c:/users ... line 22 return (np.exp(-a*x)**b) runtimewarning: overflow encountered in exp

the code far aware did not change. thanks

without knowing changed in code, hard changed between state of "working" , "not working". may changes in version of scipy used give different results: there have changes underlying implementation in curve_fit() on past few years.

but also: curve_fit() (and underlying python , fortran code uses) requires reasonably initial guesses parameters many problems work @ all. bad guesses parameters, many problems fail.

exponential decay problems seem challenging levenberg-marquardt algorithm (and implementations used curve_fit(), , require reasonable starting points. it's easy part of parameter space function evaluates zero, , changes in parameter values have no effect.

if possible, if problem involves exponential decay, helpful work in log space. is, model log(f), not f itself. problem in particular, model function exp(-a*x)**b. mean? a , bwill correlated.

in addition, may find lmfit helpful. has model class curve-fitting, using similar underlying code, allows fixing or setting bounds on of parameters. example problem (approximately):

import numpy np import matplotlib.pyplot plt import scipy.stats scipy.optimize import curve_fit  import lmfit  data= np.array([     [24, 0.176644513],     [27, 0.146382841],     [30, 0.129891534],     [33, 0.105370908],     [38, 0.077820511],     [50, 0.047407538]]) x, y = np.array([]), np.array([])  val in data:     x = np.append(x, val[0])     y = np.append(y, (val[1]/(1-val[1])))  def f(x, a, b):     print("in f:  a, b =  " , a, b)     return (np.exp(-a*x)**b)   fmod = lmfit.model(f) params = fmod.make_params(a=-0.2, b=-0.4)  # set bounds on parameters params['a'].min = -2 params['a'].max =  0 params['b'].vary = false  out = fmod.fit(y, params, x=x) print(out.fit_report())  plt.plot(x, y) plt.plot(x, out.best_fit, '--') plt.show()