design-patterns.tex

\section{Design Patterns}

TODO: define selection matrix

The literature defined selection matrix in two ways. For selecting k values from n, def1 uses an nxn square matrix with k non-zero elements in the diagonal.

Using this, one can select rows and columns as follows:

$S1 * A * S1$

The rest of the rows and columns will be zeros.

Def2 uses a kxn matrix with exactly one non-zero elements in each row.

$S2 * A * S2^T$.

Def2 changes the row/column indexes.