how to fix ValueError: operands could not be broadcast together with shapes (50,) (10,)

Question

这是数据框的样子，下面是带有方程和曲线拟合的代码：

 all # this is the dataframe 'all' 

 [        y           x            Temp      gamma         Kc         Ko
   0     -0.501864    51.040405  26.575860  4.654191  32.342013  17.434281
   1      3.317406    82.390735  26.608765  4.658023  32.457605  17.452507
   2     10.207180   146.722313  26.645362  4.662289  32.586622  17.472796
   3     16.055772   212.771464  26.698917  4.668535  32.776289  17.502520]

绘制模型数据时出现以下错误，我不明白为什么：

def model(x, Vcmax):
  return  Vcmax*((x- gamma))/((x+Kc)*(1+(O/Ko))) - Rd

def func_fit(x):
  return model(x,  
            *popt)

from scipy.optimize import curve_fit
import math

for g in all: 
  
O = 21
Rd = 1.60  

Vcmax = 40 # initial guess
gamma=g.gamma # value from dataframe, function of temperature 
Kc=g.Kc       # value from dataframe, function of temperature
Ko=g.Ko       # value from dataframe, function of temperature

#  gamma = 34
#  Kc = 89
#  Ko = 80
    
popt, pcov = curve_fit(model, g['x'], g['y'],  p0 = np.array([Vcmax]), 
                       absolute_sigma=True, maxfev=1000000)
    

##################### new data frame
new_row = { 'Vcmax': popt[0]}

results=results.append(new_row, ignore_index=True)

x = np.linspace(0,300)
b = model(x, results.iloc[0,1],) # this is the line generating the error, 

plt.plot(x, b, 'r') # plotting modelled data
plt.scatter(df.x, df.y, color='red') # plotting observations 

ValueError: operands could not be broadcast together with shapes (50,) (10,)

如果我 'fix' 将 gamma、Ko 和 Kc 的值设为常量而不是当前情况，那么我不会得到相同的错误，但是拟合非常糟糕，见图

Answer 1

我没试过，但这应该行得通：

x = np.linspace(0,300, 10)
b = model(x, results.iloc[0,1],) # this is the line generating the error,

问题是 x 和 results.iloc[0,1] 的形状不一样，但它们本应如此。将参数 10 添加到 np.linespace 应将 x 数组的形状更改为与 results.iloc[0,1]

相同的形状

how to fix ValueError: operands could not be broadcast together with shapes (50,) (10,)

how to fix ValueError: operands could not be broadcast together with shapes (50,) (10,)

python

curve-fitting