{
  "id": "transition-matrix",
  "title": "Transition Matrix",
  "lang": "en",
  "source_lang": "en",
  "summary": "A matrix encoding probabilities of movement between states, or, in linear algebra, conversion between coordinate bases.",
  "category": "mathematics",
  "subcategory": "probability-statistics",
  "type": "concept",
  "aliases": [
    "Transition probability matrix",
    "Markov matrix"
  ],
  "generated_at": "2026-10-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/transition-matrix/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/transition-matrix/",
      "json": "https://aiwiki.page/zh/transition-matrix.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A transition matrix is a matrix that represents movement between states or coordinate descriptions. In probability theory, it records the probabilities of transitions in a Markov chain. In linear algebra, the same term can denote a change-of-basis matrix, which converts coordinates between two bases of a vector space. These meanings require different assumptions: a probability transition matrix has nonnegative entries and normalized rows or columns, whereas a change-of-basis matrix must be invertible. (stat110.hsites.harvard.edu)"
    },
    {
      "heading": "Definition for a Markov chain",
      "text": "Consider a discrete-time stochastic process \\(X_0,X_1,\\ldots\\) with finite state space \\(S=\\{1,\\ldots,m\\}\\). Its Markov property means that, conditional on the current state, the next state does not depend on earlier states. For a time-homogeneous chain, the conditional probability\n\n\\[p_{ij}=\\Pr(X_{t+1}=j\\mid X_t=i)\\]\n\ndoes not depend on \\(t\\). The transition matrix is \\(P=(p_{ij})\\). Row \\(i\\) identifies the current state, and column \\(j\\) identifies the next state. Consequently,\n\n\\[p_{ij}\\geq0,\\qquad \\sum_{j=1}^{m}p_{ij}=1.\\]\n\nA matrix satisfying these conditions is called a row-stochastic matrix. Each row is a probability distribution over possible destinations, including the possibility of remaining in the same state. (stat110.hsites.harvard.edu)\n\nSome authors instead use column-stochastic matrices, whose columns sum to one. The two conventions are related by matrix transposition. The convention must be specified because it determines which side of the matrix a probability vector multiplies. A transition matrix describes conditional movement, not the initial distribution; both are needed to specify the chain's probabilistic evolution. (stat.cmu.edu)"
    },
    {
      "heading": "Evolution and multiple-step transitions",
      "text": "Under the row-stochastic convention, let \\(\\mu_t\\) be the row vector with entries \\(\\Pr(X_t=i)\\). The law of total probability gives\n\n\\[\\mu_{t+1}=\\mu_tP,\\qquad \\mu_t=\\mu_0P^t.\\]\n\nThe entry \\((P^n)_{ij}\\) is the probability of reaching state \\(j\\) after \\(n\\) steps when starting at \\(i\\). Matrix multiplication sums the probabilities of paths through intermediate states. This yields the Chapman–Kolmogorov equations,\n\n\\[P^{r+s}=P^rP^s.\\]\n\nHere \\(P^0=I\\), the identity matrix, representing zero elapsed steps. (stat.cmu.edu)\n\nFor example, consider the illustrative matrix\n\n\\[P=\\begin{pmatrix}0.8&0.2\\\\0.3&0.7\\end{pmatrix}.\\]\n\nStarting in state 1 gives \\(\\mu_0=(1,0)\\), hence \\(\\mu_1=(0.8,0.2)\\). Direct multiplication gives\n\n\\[P^2=\\begin{pmatrix}0.70&0.30\\\\0.45&0.55\\end{pmatrix}.\\]\n\nThus the two-step probability of moving from state 1 to state 2 is \\(0.30\\), obtained as \\(0.8(0.2)+0.2(0.7)\\). These are calculations for a hypothetical chain, not empirical measurements.\n\nIf transition probabilities vary with time, separate matrices \\(P_t\\) are required. Evolution then uses an ordered product, \\(\\mu_n=\\mu_0P_0P_1\\cdots P_{n-1}\\), rather than a power of one fixed matrix. (stat.cmu.edu)"
    },
    {
      "heading": "Structure and long-run behavior",
      "text": "A transition matrix can be represented by a weighted directed graph: states are vertices, and an edge \\(i\\to j\\) exists when \\(p_{ij}>0\\). The chain is irreducible if every state can reach every other state with positive probability in some number of steps. A state is absorbing if \\(p_{ii}=1\\), making departure impossible. (stat110.hsites.harvard.edu)\n\nA stationary distribution is a probability row vector \\(\\pi\\) satisfying\n\n\\[\\pi P=\\pi,\\qquad \\sum_i\\pi_i=1.\\]\n\nIt is therefore a left eigenvector associated with eigenvalue \\(1\\). For a finite irreducible chain, the stationary distribution is unique. If the chain is also aperiodic, every initial distribution converges to it, and every row of \\(P^n\\) converges to \\(\\pi\\). (probabilitycourse.com)\n\nStationarity does not itself imply convergence. The matrix\n\n\\[\\begin{pmatrix}0&1\\\\1&0\\end{pmatrix}\\]\n\nhas stationary distribution \\((1/2,1/2)\\), but a chain starting in one state alternates indefinitely. This distinction separates an invariant distribution from a limiting distribution. (cs.ox.ac.uk)"
    },
    {
      "heading": "Estimation from observations",
      "text": "For observed state sequences, let \\(N_{ij}\\) count transitions from \\(i\\) to \\(j\\). Under a time-homogeneous Markov model, conditioning on the initial state, maximum likelihood estimation gives\n\n\\[\\widehat p_{ij}=\\frac{N_{ij}}{\\sum_kN_{ik}},\\]\n\nprovided the denominator is positive. Transition counts are sufficient statistics for this conditional likelihood. If no departure from a state is observed, its transition row cannot be estimated uniquely by this formula. (stat.cmu.edu)\n\nIn a hidden Markov model, states are latent rather than directly observed. Estimation therefore uses inferred, probability-weighted transition counts, commonly within the expectation–maximization algorithm, together with a model of the observations emitted by each state. (stat.cmu.edu)"
    },
    {
      "heading": "Continuous-time transition matrices",
      "text": "For a finite, time-homogeneous continuous-time Markov chain, transitions over elapsed time \\(t\\) are described by\n\n\\[P(t)_{ij}=\\Pr(X(t)=j\\mid X(0)=i).\\]\n\nThese matrices satisfy \\(P(0)=I\\) and \\(P(s+t)=P(s)P(t)\\). Their infinitesimal generator \\(Q\\) has nonnegative off-diagonal entries and rows summing to zero. Finite-time probabilities are obtained through the matrix exponential,\n\n\\[P(t)=e^{tQ}.\\]\n\nThe generator contains transition rates, not probabilities; its diagonal entries are nonpositive. It must therefore be distinguished from \\(P(t)\\). (columbia.edu)"
    },
    {
      "heading": "The change-of-basis meaning",
      "text": "Let \\(B=(b_1,\\ldots,b_m)\\) and \\(C\\) be ordered bases of the same finite-dimensional vector space. The transition matrix from \\(B\\)-coordinates to \\(C\\)-coordinates has columns\n\n\\[T_{C\\leftarrow B}=\\bigl([b_1]_C\\ \\cdots\\ [b_m]_C\\bigr).\\]\n\nFor every vector \\(v\\),\n\n\\[[v]_C=T_{C\\leftarrow B}[v]_B.\\]\n\nIts inverse performs the reverse conversion. Unlike a probability transition matrix, it need not have nonnegative entries or normalized rows. The underlying vector remains unchanged; only its coordinate description changes. (math.hmc.edu)"
    }
  ],
  "links": [
    "matrix",
    "probability",
    "markov-chain",
    "linear-algebra",
    "stochastic-process",
    "markov-property",
    "conditional-probability",
    "probability-distribution",
    "matrix-transpose",
    "law-of-total-probability",
    "chapman-kolmogorov-equations",
    "identity-matrix",
    "directed-graph",
    "stationary-distribution",
    "eigenvalues-and-eigenvectors",
    "maximum-likelihood-estimation",
    "sufficient-statistic",
    "hidden-markov-model",
    "expectation-maximization-algorithm",
    "continuous-time-markov-chain",
    "basis-linear-algebra",
    "vector-space",
    "inverse-matrix"
  ],
  "sources": [
    {
      "title": "stat110.hsites.harvard.edu/resource/markov-chains",
      "url": "https://stat110.hsites.harvard.edu/resource/markov-chains"
    },
    {
      "title": "stat.cmu.edu/~cshalizi/dst/20/lectures/18/lecture-18.html",
      "url": "https://stat.cmu.edu/~cshalizi/dst/20/lectures/18/lecture-18.html"
    },
    {
      "title": "probabilitycourse.com/chapter11/11_2_6_stationary_and_limiting_distributions.php",
      "url": "https://www.probabilitycourse.com/chapter11/11_2_6_stationary_and_limiting_distributions.php"
    },
    {
      "title": "cs.ox.ac.uk/people/elias.koutsoupias/pc2016-17/lectures.html",
      "url": "https://www.cs.ox.ac.uk/people/elias.koutsoupias/pc2016-17/lectures.html"
    },
    {
      "title": "stat.cmu.edu/~cshalizi/dst/20/lectures/19/lecture-19.html",
      "url": "https://stat.cmu.edu/~cshalizi/dst/20/lectures/19/lecture-19.html"
    },
    {
      "title": "stat.cmu.edu/~cshalizi/dst/18/lectures/24/lecture-24.html",
      "url": "https://www.stat.cmu.edu/~cshalizi/dst/18/lectures/24/lecture-24.html"
    },
    {
      "title": "columbia.edu/~ww2040/3106F11/CTMCchapter121906.pdf",
      "url": "https://www.columbia.edu/~ww2040/3106F11/CTMCchapter121906.pdf"
    },
    {
      "title": "math.hmc.edu/calculus/hmc-mathematics-calculus-online-tutorials/linear-algebra/…",
      "url": "https://math.hmc.edu/calculus/hmc-mathematics-calculus-online-tutorials/linear-algebra/change-of-basis/"
    }
  ],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}