Created
December 5, 2022 17:40
-
-
Save JosepER/4d49958a6e015cd1419be72979f71166 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import cppyy | |
import numpy as np | |
# 'translation' of the Python function below done by ChatGPT | |
cppyy.cppdef(""" | |
#include <iostream> | |
#include <vector> | |
#include <algorithm> | |
#include <numeric> | |
double gini_cpp(const std::vector<double>& x, const std::vector<double>& w) { | |
// The rest of the code requires numpy arrays. | |
std::vector<double> sorted_x = x; | |
std::sort(sorted_x.begin(), sorted_x.end()); | |
std::vector<double> sorted_w = w; | |
std::sort(sorted_w.begin(), sorted_w.end()); | |
// Force float dtype to avoid overflows | |
std::vector<double> cumw(sorted_w.size()); | |
std::partial_sum(sorted_w.begin(), sorted_w.end(), cumw.begin()); | |
std::vector<double> cumxw(sorted_x.size()); | |
std::transform(sorted_x.begin(), sorted_x.end(), sorted_w.begin(), cumxw.begin(), std::multiplies<double>()); | |
std::partial_sum(cumxw.begin(), cumxw.end(), cumxw.begin()); | |
return (std::inner_product(cumxw.begin() + 1, cumxw.end(), cumw.begin(), 0.0) - | |
std::inner_product(cumxw.begin(), cumxw.end() - 1, cumw.begin() + 1, 0.0)) / | |
(cumxw.back() * cumw.back()); | |
} | |
""") | |
from cppyy.gbl import gini_cpp | |
# found the function here: https://stackoverflow.com/questions/48999542/more-efficient-weighted-gini-coefficient-in-python | |
def gini_py(x, w=None): | |
# The rest of the code requires numpy arrays. | |
x = np.asarray(x) | |
if w is not None: | |
w = np.asarray(w) | |
sorted_indices = np.argsort(x) | |
sorted_x = x[sorted_indices] | |
sorted_w = w[sorted_indices] | |
# Force float dtype to avoid overflows | |
cumw = np.cumsum(sorted_w, dtype=float) | |
cumxw = np.cumsum(sorted_x * sorted_w, dtype=float) | |
return (np.sum(cumxw[1:] * cumw[:-1] - cumxw[:-1] * cumw[1:]) / | |
(cumxw[-1] * cumw[-1])) | |
else: | |
sorted_x = np.sort(x) | |
n = len(x) | |
cumx = np.cumsum(sorted_x, dtype=float) | |
# The above formula, with all weights equal to 1 simplifies to: | |
return (n + 1 - 2 * np.sum(cumx) / cumx[-1]) / n | |
print(gini_cpp([6,5,8,7], [0.1,0.1,0.2,0.2])) | |
print(gini_py([6,5,8,7], [0.1,0.1,0.2,0.2])) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment