1: \begin{definition}[Matrix Partition Codes]

2: Let $P$ be a matrix satisfying \eqref{eqn2.2} and \eqref{eqn2.3}.

3: Let %$Y$ be a row basis matrix of 

4: {\color{nblue}

5: \begin{equation}

6: Y \mbox{ be a row basis matrix of } 

7: \begin{pmatrix}

8: Q_1 \\

9: \vdots \\

10: Q_s

11: \end{pmatrix}

12: \label{eqn2.18g} 

13: \end{equation}}

14: and $T$ be a matrix s.t. 

15: \begin{equation}

16: \begin{pmatrix}

17:  Y \\T

18: \end{pmatrix}

19:  \mbox{ is an invertible $n \times n$ matrix.                } 

20:  \label{eqn2.18z}

21: \end{equation}

22: % By Theorem \ref{thm2.1}

23: Then we call the compression with encoding matrices

24: \begin{equation}

25: \color{myblue}{U_1}\color{black} \begin{pmatrix}

26:   C_1 \\ G_1

27:  \end{pmatrix},

28: \cdots ,

29: \color{myblue}{U_s}\color{black}

30: \begin{pmatrix}

31:  C_s \\G_s

32: \end{pmatrix}

33: \label{eqn2.18i}

34: \end{equation}

35: % forms a compression for $\mathcal S$, 

36: a Matrix Partition Code for source $\mathcal S$,

37: where

38: {\color{mbrown}$

39:  T = \begin{pmatrix}

40:       G_1 \\ \vdots \\ G_s 

41:      \end{pmatrix}

42: $, $C_i$ are row basis matrices of $Q_i$, and $U_i$ are arbitrary invertible matrices with appropriate sizes for all $i$}.

43: \label{defM}

44: \end{definition}

45: