Skip to main content

Table 3 GeneId-Ranking With a Graph

From: A graph-search framework for associating gene identifiers with documents

(1) let Q0 be the probability distribution such that Q0(x) = 1

(2) for d = 1, ..., d max do

    • let Q d (x) = 0 for all x

    • for i = 1, ..., m do

       - sample x i according to Qd-1

       - Q d (x i ) = γQd-1(x i )

- for each edge label ℓ ∈ L (x)

          * for each node y ∈ Y(x, â„“)

             · let q xy = Pr(y|â„“, x)·Pr(â„“|x)

             · increment Q d (y) by (1 - γ)Qd-1(x i )q xy

(3) return Q d m a x MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGrbqudaWgaaWcbaGaemizaq2aaSbaaWqaaGqaciab=1gaTjab=fgaHjab=Hha4bqabaaaleqaaaaa@33B2@ (z) as an approximation to Q(z|x)

  1. An efficient approximation algorithm for computing Q(z|x), given transition probabilities Pr(y|x, â„“) and Pr(â„“|x).