我需要通过非线性最小二乘回归获得参数(kf,beta1,beta2,gamma)。错误消息是:“ ValueError:操作数不能与形状(4,7)(0,)一起广播”
我对收集到的下一个数据进行了4个实验:
我希望同时使用这四个实验来最小化参数估计的误差。
import numpy as np
from scipy.optimize import leastsq
flux = [flujo_ms, flujo_ms, flujo_ms, flujo_ms]
brfv = [[0.00694] * 7, [0.00972] * 7, [0.0139] * 7, [0.0208]*7]
fr = [fr1, fr2, fr3, fr4]
def foulingRate(parameters, flux, brfv, mlts=8.22):
kf, beta1, beta2, gamma = parameters
FR = kf * np.exp(flux * (beta1 * brfv + beta2 * mlts + gamma))
return FR
def objective(pars, yData, xData, brfv):
# it will minimize this function
err = yData - foulingRate(pars, xData, brfv)
return err
x0 = [5.6 * 10 ** -4, -2.48 * 10 ** 8, 5.1 * 10 ** 4,
2.81 * 10 ** 6] # initial values for the parameters
plsq = leastsq(objective, x0, args=(fr, flux, brfv))
print("Fitted parameters = {0}".format(plsq[0]))
我已经找到了解决方案。我使用了lmfit模块:cars9.uchicago.edu/software/python/lmfit