vectorization - trying to vectorize in python - list of variable row matrices as indices

Question

I want to perform this code in a vectorized way.

1. Turn each element of this list of lists, comboArray, into a matrix.
2. The elements of these matrices are indexes of X
3. I need to take the sum of product of the column indices of X.

For example, if comboArray = [[1, 2], [3, 4]], then I want

X[1] * X[2] + X[3] * X[4]

I'm having a lot of trouble figuring how to get rid of the for loop and do this with only products of matrices and arrays.

Here is an example:

comboArray = [
    [[1, 2]], 
    [[1, 3]], 
    [[2, 3], [1, 4]], 
    [[1, 5], [2, 4]], 
    [[3, 4], [2, 5], [1, 6]]
]

X = np.random.rand(10)
for i in range(len(comboArray)):
    gg = np.asarray(comboArray[i])
    out += np.sum( X[gg[:, 0]] * X[gg[:, 1]] )

question from:https://stackoverflow.com/questions/66054794/trying-to-vectorize-in-python-list-of-variable-row-matrices-as-indices

Answer 1

Waitting for answers