需要帮助改进此 Python function 以确定平均绝对偏差 (MAD)

Question

def what_is_mad(list_of_nums):
    xbar = what_is_mean(list_of_nums)
    for num in list_of_nums:
        value1 = list_of_nums[0]
        value2 = list_of_nums[1]
        value3 = list_of_nums[2]
        value4 = list_of_nums[3]
        test = abs(value1 - xbar) + abs(value2 - xbar) + abs(value3 - xbar) + abs(value4 - xbar)
        mad = test / xbar
    return int(mad)

what_is_mad(2, 2, 4, 4)

I'm trying to build my own purely Python function for determining Mean Absolute Deviation (MAD) and would like some recommendations on how I can enhance this code. It is currently limited to four values, as seen above, but I'd like to be able to have the function ingest any number of values. Thanks in advance.

Answer 1

The way(formula) of calculating the Mean absolute deviation is wrong in your code. You should calculate all your absolute deviations, sum them up and divide by the number of elements in the list_of_sums, not by the mean. If you mean pure python as not using any libraries this code will do.

xbar = sum(list_of_nums)/len(list_of_nums)
dev = 0
for num in list_of_nums:
    dev = dev + abs(num - xbar)
mad = dev/len(list_of_nums)
print(mad)

You can use list comprehension and write the code in less number of lines. But it would create another list which will cost space to store it, hence not optimized.

Otherwise you can use numpy, scipy or pandas and they have inbuilt functions to calculate MAD

Answer 2

xbar = mean(list_of_nums)
test=0
for num in list of nums:
    test = test + abs(num - xbar)
mad = test/xbar
return int(mad)

This loops through the list and sums up the absolute deviations one at a time, then does the mad calculation.

Answer 3

In terms of an optimized solution, pandas has a method for this.

mad_in = [1,2,3,4...]

mad_out = pandas.Series(mad_in).mad()

Answer 4

This can be accomplished in a couple lines using python's built-in sum operation and list comprehension

def what_is_mad(*list_of_nums):
    count = len(list_of_nums)
    mean = sum(list_of_nums) / count
    mad = sum((abs(num - mean) for num in list_of_nums)) / count
    return mad