Getting the standard deviation of a Eigen::vector

Question

I need to have standard deviation of a vector in Eigen library. I have not found it. So i tried it:

Eigen::VectorXd ys(5);
 ys << 1, 2, 3, 4, 5;            

double std_dev = sqrt((ys - ys.mean()).square().sum() / (ys.size() - 1)); // Error with minus sign (ys-ys.mean())

But getting error.

error:

Severity    Code    Description Project File    Line    Suppression State
Error (active)  E0349   no operator "-" matches these operands  

Answer 1

An Eigen::VectorXd is defined as typedef Matrix<double, Dynamic, 1> VectorXd; so it is a special form of an Eigen::Matrix. You are trying to subtract a scalar ys.mean() from a vector ys which is an coefficient-wise operation. The Eigen::Matrix class is not intended to be used with coefficient-wise operations but for linear algebra. For performing coefficient-wise operations Eigen has the Eigen::Array class.

Therefore it is sufficient to convert your Eigen::Matrix ys to an Eigen::Array for your formula to work:

double const std_dev = sqrt((ys.array() - ys.mean()).square().sum() / (ys.size() - 1));