Simpler way ?

[lpatiny]

Could we have another method that calculates this internally ?

const diff = xValueExperimental.map((experimental) => {
  return xValuePredicted.map((predicted) => {
    return Math.abs(predicted - experimental);
  });
});

I don't know about a correct name but it would be something like:

abcd(experimental, predicted, {})

And one of the options could be a callback like
distance: (a,b) => Math.abs(a-b)

[jobo322]

There is a PR on ml-spectra-processing adding the function xCostMatrix with it is possible to generate a cost matrix, it accepts a function as an option to generate the cost matrix as you want.

So with this we will have

import { xcostMatrix } from 'ml-spectra-processing';

const diff = xCostMatrix(experimental, predicted, { 
   fct: (a, b) => Math.abs(a - b)
});

by default the function used by xCostMatrix is the absolute value of the difference