使用Boost Accumulators和Eigen :: Vector类型 [英] Using Boost Accumulators with Eigen::Vector types

问题描述

我在将 Eigen :: VectorXd 类型与Boost累加器库结合时遇到一些问题：

I am having some problems combining Eigen::VectorXd types with the Boost accumulator library:

#include <iostream>
#include <Eigen/Core>
#include <boost/accumulators/accumulators.hpp>
#include <boost/accumulators/statistics/stats.hpp>
#include <boost/accumulators/statistics/mean.hpp>

using namespace boost::accumulators;
using namespace Eigen;

int main()
{
   Vector2f a(1.0, 2.0), b(3.0, 10.0);

   accumulator_set<Vector2f, stats<tag::mean> > acc(Vector2f::Zero());

   acc(a);
   acc(b);

   std::cout << mean(acc) << std::endl;
   std::cout << ((a+b)/2.0) << std::endl;

   return 0;
}

在我的系统上，这会产生：

On my system this produces:

4.41629e-39
0
2
6

所以当直接计算很好时（特征向量支持所有通常的数值运算符）Boost accumulators在运行时失败，没有错误。

So while direct computation is fine (Eigen vectors support all of the usual numerical operators) Boost accumulators fail at runtime without an error.

推荐答案

用户定义的类型需要specialize std :: numeric_limits。
请参阅 https://svn.boost.org/trac/boost/ticket/ 5491

User defined type need specialize std::numeric_limits. see https://svn.boost.org/trac/boost/ticket/5491

这篇关于使用Boost Accumulators和Eigen :: Vector类型的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！