Tips: Centering an array

It's very easy to center a data set using high level template expressions and statistical functors.

Consider the following code and its output:

Example:

Output:

#include "STKpp.h" using namespace STK; int main(int argc, char *argv[]) { CArrayXX A(100, 5); Law::Normal law(1,2); A.rand(law); // call column statistical functions stk_cout << _T("min(A) =") << Stat::min(A); stk_cout << _T("max(A) =") << Stat::max(A); stk_cout << _T("max(A.abs()) =") << Stat::max(A.abs()); stk_cout << _T("mean(A) =") << Stat::mean(A); // center the array stk_cout << _T("\nCentering...\n\n"); A -= Const::VectorX(100) * Stat::mean(A); // call column statistical functions with all the columns of A centered stk_cout << _T("min(A) =") << Stat::min(A); stk_cout << _T("max(A) =") << Stat::max(A); stk_cout << _T("max(A.abs()) =") << Stat::max(A.abs()); stk_cout << _T("mean(A) =") << Stat::mean(A); }

min(A) =-4.66112 -3.1092 -2.76891 -4.65294 -5.47609 max(A) =4.17889 5.89107 5.08115 6.39462 5.17813 max(A.abs()) =4.66112 5.89107 5.08115 6.39462 5.47609 mean(A) =0.963606 1.26585 1.26198 1.05056 0.723175 Centering... min(A) =-5.62472 -4.37505 -4.0309 -5.7035 -6.19926 max(A) =3.21529 4.62522 3.81916 5.34406 4.45495 max(A.abs()) =5.62472 4.62522 4.0309 5.7035 6.19926 mean(A) =2.57572e-16 -3.93019e-16 2.13163e-16 -3.4639e-16 -1.15463e-16

The expression

Const::Vector<Real>(100) * mean(A)

represents the matrix multiplication of a column vector of 1 with 100 rows and of row vector with the mean of each column of A.

Note

For each column of the array A we can get the maximal value in absolute value using max(A.abs()). It is possible to use functors mixed with unary or binary operators.