Contents

Parity-check matrix

Formally, a parity check matrix H of a linear code C is a generator matrix of the dual code, C^⊥. This means that a codeword c is in Cif and only if the matrix-vector product Hc^⊤ = 0 (some authors^[1] would write this in an equivalent form, cH^⊤ = 0.)

The rows of a parity check matrix are the coefficients of the parity check equations.^[2] That is, they show how linear combinations of certain digits (components) of each codeword equal zero. For example, the parity check matrix

${\displaystyle H=\left[{\begin{array}{cccc}0&0&1&1\\1&1&0&0\end{array}}\right]}$,

compactly represents the parity check equations,

${\displaystyle {\begin{aligned}c_{3}+c_{4}&=0\\c_{1}+c_{2}&=0\end{aligned}}}$,

that must be satisfied for the vector ${\displaystyle (c_{1},c_{2},c_{3},c_{4})}$ to be a codeword of C.

From the definition of the parity-check matrix it directly follows the minimum distance of the code is the minimum number d such that every d - 1 columns of a parity-check matrix H are linearly independent while there exist d columns of H that are linearly dependent.

The parity check matrix for a given code can be derived from its generator matrix (and vice versa).^[3] If the generator matrix for an [n,k]-code is in standard form

${\displaystyle G={\begin{bmatrix}I_{k}|P\end{bmatrix}}}$,

then the parity check matrix is given by

${\displaystyle H={\begin{bmatrix}-P^{\top }|I_{n-k}\end{bmatrix}}}$,

because

${\displaystyle GH^{\top }=P-P=0}$.

Negation is performed in the finite field F_q. Note that if the characteristic of the underlying field is 2 (i.e., 1 + 1 = 0 in that field), as in binary codes, then -P = P, so the negation is unnecessary.

For example, if a binary code has the generator matrix

${\displaystyle G=\left[{\begin{array}{cc|ccc}1&0&1&0&1\\0&1&1&1&0\\\end{array}}\right]}$,

then its parity check matrix is

${\displaystyle H=\left[{\begin{array}{cc|ccc}1&1&1&0&0\\0&1&0&1&0\\1&0&0&0&1\\\end{array}}\right]}$.

It can be verified that G is a ${\displaystyle k\times n}$ matrix, while H is a ${\displaystyle (n-k)\times n}$ matrix.

For any (row) vector x of the ambient vector space, s = Hx^⊤ is called the syndromeofx. The vector x is a codeword if and only if s = 0. The calculation of syndromes is the basis for the syndrome decoding algorithm.^[4]

• Hamming code

• Hill, Raymond (1986). A first course in coding theory. Oxford Applied Mathematics and Computing Science Series. Oxford University Press. pp. 69. ISBN 0-19-853803-0.
• Pless, Vera (1998), Introduction to the Theory of Error-Correcting Codes (3rd ed.), Wiley Interscience, ISBN 0-471-19047-0
• Roman, Steven (1992), Coding and Information Theory, GTM, vol. 134, Springer-Verlag, ISBN 0-387-97812-7
• J.H. van Lint (1992). Introduction to Coding Theory. GTM. Vol. 86 (2nd ed.). Springer-Verlag. pp. 34. ISBN 3-540-54894-7.