Hi Martijn, This indeed sounds too domain-specific for NumPy; SciPy would be a better fit. Also note that you can already rewrite this in terms of scipy.special.logsumexp:
c = np.log(10) / 10 log_mean = scipy.special.logsumexp(c * x, b=1 / x.size) / c Hope that helps, Joren _______________________________________________ NumPy-Discussion mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3//lists/numpy-discussion.python.org Member address: [email protected]
