函数中使用数组出错：numpy.ufunc与float不支持**或pow()操作

Hey there, let's fix this error and get your curve fitting working properly! Here's what's going wrong and how to correct it step by step:

1. 核心错误：误用numpy的exp函数 #

The error you're seeing comes directly from this line in your Radio function:

return (n*(exp**(-(t/tao)))) + b

exp is a numpy universal function (ufunc)—it's not a numerical value, so you can't use the ** (power operator) on it. The correct way to calculate the exponential of -t/tao is to pass that value as an argument to the exp() function, like this:

return n * exp(-t / tao) + b

2. 函数参数顺序问题（潜在的拟合失败隐患） #

Looking at how you're using curve_fit, your function's parameter order is backwards. curve_fit expects the first parameter to be the independent variable (in your case, t or T), followed by the parameters you want to fit. Your current definition has n (the dependent data) as the first parameter, which will cause curve_fit to misinterpret your inputs.

Fix the function definition to match the correct order. For example, if you're fitting the model N(t) = N0 * exp(-t/tao) + b, define it as:

def Radio(t, n0, tao, b):
    return n0 * exp(-t / tao) + b

3. 语法错误：无效的猜测值元组 #

Your guesses = (1,1...) line has invalid syntax—you can't use ... in a tuple like that. Replace it with concrete initial guess values for your parameters. For example, if you expect n0 around 100, tao around 5, and b around 0, write:

guesses = (100, 5, 0)

修正后的完整代码片段 #

Here's how your corrected code might look (adjust the guess values based on your actual data):

import numpy as np
import matplotlib.pyplot as plt
from numpy import sqrt, exp, log
from scipy import linalg
from scipy.optimize import curve_fit

data1 = np.loadtxt('decay1.txt', float, skiprows=1)
t = data1[:,0]
n = data1[:,1]
data2 = np.loadtxt('decay2.txt', float, skiprows=1)
T = data2[:,0]
N = data2[:,1]

# Corrected function definition
def Radio(t, n0, tao, b):
    return n0 * exp(-t / tao) + b

# Valid initial guesses (adjust these to match your data's expected values)
guesses = (100, 5, 0)

# Example curve fit call for the first dataset
params, covariance = curve_fit(Radio, t, n, p0=guesses)
n_fit = Radio(t, *params)

# Plot to check the fit
plt.scatter(t, n, label='Raw data')
plt.plot(t, n_fit, 'r-', label='Fitted curve')
plt.xlabel('Time')
plt.ylabel('Count')
plt.legend()
plt.show()

为什么这些修改有效？ #

• Fixing the exp() syntax removes the invalid operation between a ufunc and a float, which was causing your original error.
• Reordering the function parameters ensures curve_fit correctly maps your independent variable (t) and dependent data (n) to the model.
• Using a valid tuple for guesses avoids syntax errors and helps curve_fit converge to the correct parameters faster.

内容的提问来源于stack exchange，提问作者Brandon SeedlessBananas Mc-Wil