{
  "id": "householder-transformation",
  "title": "Householder Transformation",
  "lang": "en",
  "source_lang": "en",
  "summary": "A Householder transformation is a reflection across a hyperplane, used to eliminate vector components and construct stable matrix factorizations.",
  "category": "mathematics",
  "subcategory": "applied",
  "type": "concept",
  "aliases": [
    "Householder reflection",
    "Householder reflector",
    "Householder matrix",
    "elementary reflector"
  ],
  "generated_at": "2026-10-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/householder-transformation/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/householder-transformation/",
      "json": "https://aiwiki.page/zh/householder-transformation.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A Householder transformation is a linear map that reflects vectors across a hyperplane through the origin. In real Euclidean space, it is represented by a symmetric orthogonal matrix differing from the identity by a rank-one matrix. Its principal computational use is to eliminate selected components of a vector while preserving lengths, making it a fundamental tool in numerical linear algebra, especially for QR decomposition and reductions preceding eigenvalue and singular-value computations. The transformation is named after Alston S. Householder, whose 1958 paper described its use in matrix triangularization. (web.stanford.edu)"
    },
    {
      "heading": "Definition and geometric interpretation",
      "text": "For a nonzero vector \\(v\\in\\mathbb{R}^n\\), the Householder matrix is\n\n\\[H=I-2\\frac{vv^T}{v^Tv},\\]\n\nwhere \\(I\\) is the identity matrix and \\(v^T\\) denotes the transpose of \\(v\\). Equivalently, with the unit vector \\(u=v/\\|v\\|_2\\),\n\n\\[H=I-2uu^T.\\]\n\nThis is the real reflection form of the elementary-reflector representation used in matrix algorithms. (web.stanford.edu)\n\nThe geometry follows directly from the formula. Decompose any vector \\(x\\) as\n\n\\[x=x_\\perp+x_\\parallel, \\qquad x_\\parallel=u(u^Tx), \\qquad u^Tx_\\perp=0.\\]\n\nThen\n\n\\[Hx=x_\\perp-x_\\parallel.\\]\n\nThus, \\(H\\) reverses the component parallel to \\(v\\) and leaves its orthogonal complement unchanged. The fixed hyperplane is \\(v^\\perp=\\{x:v^Tx=0\\}\\). In terms of orthogonal projection, \\(H=I-2P\\), where \\(P=uu^T\\) projects onto the line spanned by \\(v\\). (web.stanford.edu)"
    },
    {
      "heading": "Algebraic properties",
      "text": "Direct calculation gives\n\n\\[H^T=H, \\qquad H^2=I, \\qquad H^TH=I.\\]\n\nConsequently, the transformation is symmetric, orthogonal, and its own inverse. Orthogonality implies preservation of the inner product:\n\n\\[(Hx)^T(Hy)=x^Ty.\\]\n\nIt therefore preserves the Euclidean norm, angles, and distances. These identities explain why reflections are useful for transforming numerical problems without amplifying errors simply through the transformation itself. (web.stanford.edu)\n\nFurther consequences of the defining formula include:\n\n- \\(Hv=-v\\), so \\(v\\) is an eigenvector with eigenvalue \\(-1\\).\n- Every vector perpendicular to \\(v\\) has eigenvalue \\(+1\\).\n- The determinant is \\(-1\\).\n- \\(I-H\\) has rank one.\n- Replacing \\(v\\) by any nonzero scalar multiple produces the same \\(H\\).\n\nFor \\(n>1\\), the eigenvalues are therefore \\(-1\\) once and \\(+1\\) with multiplicity \\(n-1\\). A product of Householder matrices remains orthogonal, although it need not itself be a single reflection. (web.stanford.edu)"
    },
    {
      "heading": "Constructing a reflector to eliminate components",
      "text": "Given a nonzero \\(x\\in\\mathbb{R}^n\\), a common objective is to construct \\(H\\) such that\n\n\\[Hx=\\alpha e_1,\\]\n\nwhere \\(e_1=(1,0,\\ldots,0)^T\\). Length preservation requires \\(|\\alpha|=\\|x\\|_2\\). In exact arithmetic, setting\n\n\\[v=x-\\alpha e_1\\]\n\nand using the defining Householder formula achieves this whenever \\(v\\ne0\\). All components except the first are then annihilated in one operation. (web.stanford.edu)\n\nFor computation in floating-point arithmetic, the conventional choice is\n\n\\[\\alpha=-s\\|x\\|_2, \\qquad s= \\begin{cases} 1,&x_1\\ge0,\\\\ -1,&x_1<0. \\end{cases}\\]\n\nThen\n\n\\[v=x+s\\|x\\|_2e_1.\\]\n\nThe first component of \\(v\\) is formed by adding quantities of the same sign, avoiding the cancellation that could occur with the opposite choice when \\(x\\) is nearly parallel to \\(e_1\\). Practical implementations also rescale intermediate quantities when necessary to avoid underflow or overflow. If the components to be eliminated are already zero, software may use the identity transformation rather than perform a reflection; this is a computational convention, not a nontrivial geometric reflection. (netlib.org)\n\nExample\n\nLet\n\n\\[x=\\begin{pmatrix}3\\\\4\\end{pmatrix}.\\]\n\nSince \\(\\|x\\|_2=5\\), choose \\(\\alpha=-5\\) and \\(v=(8,4)^T\\). Substitution gives\n\n\\[H= I-\\frac{2}{80} \\begin{pmatrix} 64&32\\\\ 32&16 \\end{pmatrix} = \\begin{pmatrix} -\\frac35&-\\frac45\\\\ -\\frac45&\\frac35 \\end{pmatrix}.\\]\n\nMultiplication verifies that \\(Hx=(-5,0)^T\\): the second component is eliminated and the length remains \\(5\\)."
    },
    {
      "heading": "Efficient application and storage",
      "text": "Explicitly forming the dense matrix \\(H\\) is usually unnecessary. With\n\n\\[H=I-\\tau vv^T, \\qquad \\tau=\\frac{2}{v^Tv},\\]\n\nits action on a vector is\n\n\\[Hx=x-\\tau v(v^Tx).\\]\n\nThis requires a dot product and a scaled vector update, taking \\(O(n)\\) operations rather than the \\(O(n^2)\\) work of a general dense matrix–vector multiplication. For a matrix \\(A\\), multiplication from either side can likewise be expressed as a rank-one update:\n\n\\[HA=A-\\tau v(v^TA), \\qquad AH=A-\\tau(Av)v^T.\\]\n\nThese are direct computational consequences of the elementary-reflector representation. (netlib.org)\n\nSoftware often normalizes the stored reflector vector so that its first active component is \\(1\\), retaining only its remaining components and the scalar \\(\\tau\\). A sequence of reflectors can represent an orthogonal matrix implicitly, allowing multiplication by that matrix or its transpose without constructing it explicitly. (netlib.org)\n\nFor blocked computations, a product can be represented as\n\n\\[H_1H_2\\cdots H_b=I-VTV^T,\\]\n\nwhere the columns of \\(V\\) contain reflector vectors and \\(T\\) is upper triangular. This compact WY representation permits application through matrix–matrix operations, improving the use of high-performance numerical kernels. The product is sometimes called a block reflector, but it generally is not symmetric or involutory like an individual reflection. (netlib.org)"
    },
    {
      "heading": "QR factorization and least-squares problems",
      "text": "For a real \\(m\\times n\\) matrix \\(A\\), with \\(m\\ge n\\), Householder QR successively eliminates entries below the diagonal. At step \\(k\\), a reflector acts only on rows \\(k,\\ldots,m\\), leaving previously established zeros undisturbed. If the embedded reflectors are \\(H_1,\\ldots,H_p\\), then\n\n\\[R=H_p\\cdots H_2H_1A, \\qquad Q=H_1H_2\\cdots H_p,\\]\n\nso that \\(A=QR\\), with \\(Q\\) orthogonal and \\(R\\) upper trapezoidal. A reduced factorization retains only the first \\(n\\) columns of \\(Q\\) and the corresponding square triangular factor. (web.stanford.edu)\n\nThe connection to linear least squares follows from norm preservation. Writing\n\n\\[Q^Tb= \\begin{pmatrix}c_1\\\\c_2\\end{pmatrix}, \\qquad R= \\begin{pmatrix}R_1\\\\0\\end{pmatrix},\\]\n\ngives\n\n\\[\\|Ax-b\\|_2^2 = \\|R_1x-c_1\\|_2^2+\\|c_2\\|_2^2.\\]\n\nWhen \\(A\\) has full column rank, the minimizer is obtained by solving the triangular system \\(R_1x=c_1\\). The identity illustrates why orthogonal triangularization can solve least-squares problems without explicitly forming \\(A^TA\\). (web.stanford.edu)"
    },
    {
      "heading": "Eigenvalue and singular-value computations",
      "text": "Householder transformations also provide structured reductions before iterative spectral computations:\n\n- A real symmetric matrix can be reduced to tridiagonal form by orthogonal similarity transformations.\n- A general square matrix can be reduced to upper Hessenberg form, with zeros below the first subdiagonal, as a preliminary step for algorithms such as the QR algorithm.\n- A rectangular matrix can be reduced to bidiagonal form using transformations from both sides, preparing it for singular value decomposition.\n\nThese reductions preserve the relevant spectral information: similarity preserves eigenvalues, while orthogonal multiplication from the left and right preserves singular values. A general nonsymmetric matrix is not, in general, reducible to tridiagonal form by the same orthogonal-similarity procedure used for symmetric matrices. (netlib.org)"
    },
    {
      "heading": "Complex transformations",
      "text": "For vectors over the complex numbers, the direct reflection analogue is\n\n\\[H=I-2\\frac{vv^*}{v^*v},\\]\n\nwhere \\(v^*\\) is the conjugate transpose. This matrix is Hermitian and unitary, and satisfies \\(H^2=I\\). (netlib.org)\n\nNumerical libraries also use a broader complex elementary-reflector convention,\n\n\\[H=I-\\tau vv^*,\\]\n\nwith complex \\(\\tau\\). Unitarity requires\n\n\\[\\tau+\\overline{\\tau} = |\\tau|^2(v^*v).\\]\n\nSuch an \\(H\\) need not be Hermitian or involutory. For example, LAPACK’s complex reflector generator constructs \\(H\\) so that \\(H^*\\) maps the input vector to a real leading component followed by zeros. Thus, in complex computations, “Householder reflector” may denote a unitary rank-one modification rather than a literal order-two reflection. (netlib.org)"
    },
    {
      "heading": "Numerical stability and limitations",
      "text": "Properly implemented Householder QR has strong numerical stability properties. In backward-error terms, its computed factors can be interpreted as the factors of a nearby input matrix, with small normwise perturbation and a numerically orthogonal computed \\(Q\\). This does not ensure small relative errors in every individual matrix entry or in every solution derived from the factorization. (netlib.org)\n\nNor do orthogonal transformations cure an ill-conditioned problem. In exact arithmetic, they preserve the spectral-norm condition number of an invertible matrix. Accurate transformation and sensitivity of the original problem are therefore separate issues. (web.stanford.edu)\n\nHouseholder reflectors eliminate an entire group of components at once, making them well suited to dense matrix reductions. Givens rotations, by contrast, act on two coordinates at a time and provide more localized transformations. Householder’s original triangularization paper explicitly compared the two approaches, emphasizing the reduction in arithmetic work achieved by replacing sequences of plane rotations with reflections. (web.stanford.edu)"
    }
  ],
  "links": [
    "linear-map",
    "hyperplane",
    "euclidean-space",
    "orthogonal-matrix",
    "numerical-linear-algebra",
    "qr-decomposition",
    "identity-matrix",
    "matrix-transpose",
    "orthogonal-complement",
    "orthogonal-projection",
    "inner-product",
    "norm-mathematics",
    "determinant",
    "matrix-rank",
    "eigenvalues-and-eigenvectors",
    "floating-point-arithmetic",
    "matrix",
    "ordinary-least-squares",
    "system-of-linear-equations",
    "matrix-similarity",
    "qr-algorithm",
    "singular-value-decomposition",
    "complex-number",
    "conjugate-transpose",
    "unitary-matrix",
    "numerical-stability",
    "condition-number"
  ],
  "sources": [
    {
      "title": "Unitary Triangularization of a Nonsymmetric Matrix",
      "url": "https://web.stanford.edu/class/cme324/classics/householder.pdf"
    },
    {
      "title": "Representation of Orthogonal or Unitary Matrices",
      "url": "https://www.netlib.org/lapack/lug/node128.html"
    },
    {
      "title": "Non-Negative Diagonals and High Performance on",
      "url": "https://www.netlib.org/lapack/lawnspdf/lawn203.pdf"
    },
    {
      "title": "LAPACK: SRC/dlarfgp.f Source File",
      "url": "https://www.netlib.org/lapack/explore-html/d1/d44/dlarfgp_8f_source.html"
    },
    {
      "title": "LAPACK: zlarfg",
      "url": "https://www.netlib.org/lapack/explore-html/d8/d0d/group__larfg_gad96b19d25ca0cd3fc34292a8bfaf6e4c.html"
    },
    {
      "title": "LAPACK: larf: apply Householder reflector",
      "url": "https://netlib.org/lapack/explore-html/d2/d97/group__larf.html"
    },
    {
      "title": "QR Factorization",
      "url": "https://netlib.org/lapack/lug/node69.html"
    },
    {
      "title": "LAPACK: clarft",
      "url": "https://netlib.org/lapack/explore-html/d7/d0d/group__larft_ga1142eb8946252fe7ff50ada11ee56ad9.html"
    },
    {
      "title": "LAPACK Working Note #2: Block Algorithms for Reducing Symmetric and General Matrices to Tridiagonal, Bidiagonal and Hessenberg Forms",
      "url": "https://www.netlib.org/lapack/lawnspdf/lawn02.pdf"
    },
    {
      "title": "LAPACK Users’ Guide: Block Algorithms for Eigenvalue Problems",
      "url": "https://www.netlib.org/lapack/lug_old/node69.html"
    },
    {
      "title": "LAPACK Working Note #176",
      "url": "https://www.netlib.org/lapack/lawnspdf/lawn176.pdf"
    }
  ],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}