卡方检验带误差线的数据

Question

Suppose I have two arrays, A1 and A2 :

A1 = [1,2,2,2,3,3,7]
A2 = [0.5,1,1.5,2,7,2,6.9]

Suppose I also have the error bar data for A1 :

err_bars = [0.1,0.6,0.9,0.2,0.01,0.8,0.1]

I want to see if there is a way to determine if the data in A2 fits with that in A1 .

Is there any way to calculate the Chi-squared, F-test and p -value considering the error bars in the data?

So far I could not find any Chi-squared function in Python that considers the error bars of the data.

Answer 1

Try this:

   # obs -- observed value
    # exp -- expected value

def chisqr(obs, exp, error):
    chisqr = 0
    for i in range(len(obs)):
        chisqr = chisqr + ((obs[i]-exp[i])**2)/(error[i]**2)
    return chisqr

A1 = [1,2,2,2,3,3,7]
A2 = [0.5,1,1.5,2,7,2,6.9]
err_bars = [0.1,0.6,0.9,0.2,0.01,0.8,0.1]
print chisqr(A1, A2, err_bars)

Answer 2

If your working with numerical data, chances are you should be using the numpy package to do most of your calculations. For anything more than a small calculation its significantly faster, and its almost always easier to read (A good summery why is here ). Switching to to numpy lets you write this as

A1 = np.array([1, 2, 2, 2, 3, 3, 7])
A2 = np.array([0.5, 1, 1.5, 2, 7, 2, 6.9])
err_bars = np.array([0.1, 0.6, 0.9, 0.2, 0.01, 0.8, 0.1])

def chisq(obs, exp, error):
    return np.sum((obs - exp) ** 2 / (error ** 2))

print(chisq(A1, A2, err_bars))