python - 运行时警告 : divide by zero encountered in true_divide W = 1/sigma**2/s_sq symfit

Question

我试过运行这段代码，但无论我如何简化它，我总是遇到同样的错误。

/home/runner/.site-packages/symfit/core/fit.py:1046: RuntimeWarning:divide by zero encountered in true_divide W = 1/sigma**2/s_sq[:,np.newaxis]

/home/runner/.site-packages/symfit/core/fit.py:1783:> RuntimeWarning: invalid value encountered in double_scalars return 1 SS_res/SS_tot

帮助将不胜感激。

x1, ya = sf.variables('x1, ya')
I1, I2, I3, A, B, C, D = sf.parameters('I1, I2, I3, A, B, C, D')


I1.value = 46.483
I2.value = 5.916
I3.value = 21.90
A.value = -3.828*10**(-5)
B.value = 0
C.value = 0
D.value = 0

# Making the equation
ya = (A*x1**3 + B*x1**2 + C*x1 + D + (q_subm*x1**4)/(24*EI))  #q_sub and EI are constants

#model = Model({y: Piecewise((ya, x1 <= I1), (yb, x2 <= I2), (yc, x3 <= I3))})
model = Model({ya})

# As a constraint, we want cable to be at 0 at start and and at "-p" height at x=60
# also no angle in pipe ate those points
constraints = [
    Eq(ya.subs({x1: 0}), 0),
    Eq(ya.subs({x1: 60}), -p),
    Eq(ya.diff(x1).subs({x1: 0}), 0),
    Eq(ya.diff(x1).subs({x1: 60}), 0)
]


x1data = np.linspace(0, 60., 60)
y1data = model(x1=x1data, A = -3.828*10**(-5), B = 0, C = 0, D = 0)
np.random.seed(2)
y1data = np.random.normal(y1data, 0.005) 

plt.plot([60], [0.4], 'ro')
plt.scatter(x1data, y1data)
plt.savefig('plot.png')

print ('Done plotting fig')

#fit = Fit(model, x=xdata, y=ydata, constraints=constraints)
fit = Fit(model, x1=x1data, constraints=constraints)

print ('Done fitting model')
fit_result = fit.execute()
#print(fit_result)

Answer 1

我对您的代码有一些评论，也许其中之一可以解决问题。

• model = Model({ya}) 错误地尝试从集合中构建模型，因为 {} 构建了集合。尝试使用 model = Model({y: ya}) 或 model = Model(ya) 代替。 (我会推荐第一个)
• 注释行 fit = Fit(model, x=x1data , y=y1data, constraints=constraints) 基本上是正确的，请注意我更改了数据数组的名称以包含 1。上面示例中的行无法工作，因为您没有提供 ydata。
• q_subm 和EI 的类型 是什么？只要它们是标准的 Python 数字类型就没问题，但如果它们是奇异的东西，则可能会导致问题。
• 请记住，您在问题中列出的“错误”是警告，只要结果正确，在拟合过程中出现一些警告就没有问题。

我希望这能解决问题，如果不能让我知道。

x1, y = sf.variables('x1, y')
I1, I2, I3, A, B, C, D = sf.parameters('I1, I2, I3, A, B, C, D')


I1.value = 46.483
I2.value = 5.916
I3.value = 21.90
A.value = -3.828*10**(-5)
B.value = 0
C.value = 0
D.value = 0

# Making the equation
ya = (A*x1**3 + B*x1**2 + C*x1 + D + (q_subm*x1**4)/(24*EI))  #q_sub and EI are constants
model = Model({y: ya})

# As a constraint, we want cable to be at 0 at start and and at "-p" height at x=60
# also no angle in pipe ate those points
constraints = [
    Eq(ya.subs({x1: 0}), 0),
    Eq(ya.subs({x1: 60}), -p),
    Eq(ya.diff(x1).subs({x1: 0}), 0),
    Eq(ya.diff(x1).subs({x1: 60}), 0)
]


x1data = np.linspace(0, 60., 60)
y1data = model(x1=x1data, A=-3.828*10**(-5), B=0, C=0, D=0)
np.random.seed(2)
y1data = np.random.normal(y1data, 0.005) 

plt.plot([60], [0.4], 'ro')
plt.scatter(x1data, y1data)
plt.savefig('plot.png')

print ('Done plotting fig')

fit = Fit(model, x=x1data, y=y1data, constraints=constraints)
fit_result = fit.execute()

print ('Done fitting model')
print(fit_result)