Scribd Logo
Upload

Welcome to Scribd!

  • 53 views

    Cost Matrix

    Uploaded by

    sajjadmir
    The document describes calculating the cost matrix for segmenting genomic data from multiple arrays. The cost matrix Gkm is the sum of squared residuals for a segment from position m to m+k-1. It can be calculated efficiently in O(NJ) time by pre-computing vectors q and r of summed squares and sums of the data, then using cumulative sums c and d to calculate the cost matrix elements per equation (10). This avoids the O(N2) complexity of directly calculating each element.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    Cost Matrix

    Uploaded by

    sajjadmir
    53 views2 pages
    The document describes calculating the cost matrix for segmenting genomic data from multiple arrays. The cost matrix Gkm is the sum of squared residuals for a segment from position m to m+k-1. It can be calculated efficiently in O(NJ) time by pre-computing vectors q and r of summed squares and sums of the data, then using cumulative sums c and d to calculate the cost matrix elements per equation (10). This avoids the O(N2) complexity of directly calculating each element.

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document describes calculating the cost matrix for segmenting genomic data from multiple arrays. The cost matrix Gkm is the sum of squared residuals for a segment from position m to m+k-1. It can be calculated efficiently in O(NJ) time by pre-computing vectors q and r of summed squares and sums of the data, then using cumulative sums c and d to calculate the cost matrix elements per equation (10). This avoids the O(N2) complexity of directly calculating each element.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    53 views2 pages

    Cost Matrix

    Uploaded by

    sajjadmir
    The document describes calculating the cost matrix for segmenting genomic data from multiple arrays. The cost matrix Gkm is the sum of squared residuals for a segment from position m to m+k-1. It can be calculated efficiently in O(NJ) time by pre-computing vectors q and r of summed squares and sums of the data, then using cumulative sums c and d to calculate the cost matrix elements per equation (10). This avoids the O(N2) complexity of directly calculating each element.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 2

    Calculation of the cost matrix

    Wolfgang Huber April 13, 2011

    Problem statement and denitions

    Let ynj be the data value at position (genomic coordinate) n = 1, . . . , N for replicate array j = 1, . . . , J. Hence we have J arrays and sequences of length N . The goal of this note is to describe an O(N J) algorithm to calculate the cost matrix of a piecewise linear model for the segmentation of the (1, . . . , N ) axis. It is implemented in the function costMatrix in the package tilingArray. The cost matrix is the input for a dynamic programming algorithm that nds the optimal (least squares) segmentation. The cost matrix Gkm is the sum of squared residuals for a segment from m to m + k 1 (i. e. including m + k 1 but excluding m + k),
    J m+k1

    Gkm :=
    j=1 n=m

    (ynj km )2

    (1)

    where 1 m m + k 1 N and km is the mean of that segment, km = 1 Jk


    J m+k1

    ynj .
    j=1 n=m

    (2)

    Sidenote: a perhaps more straightforward denition of a cost matrix would be Gm m = G(m m) m , the sum of squared residuals for a segment from m to m 1. I use version (1) because it makes it easier to use the condition of maximum segment length (k <= kmax ), which I need to get the algorithm from complexity O(N 2 ) to O(N ).

    Algebra
    J m+k1

    Gkm =
    j=1 n=m 2 ynj n,j

    (ynj km )2 1 yn j Jk n ,j
    2

    (3)

    (4)

    =
    n

    1 qn Jk

    rn
    n

    (5)

    with qn :=
    j 2 ynj

    (6) (7)

    rn :=
    j

    ynj

    If y is an N J matrix, then the N -vectors q and r can be obtained by q = rowSums(y*y) r = rowSums(y) Now dene

    c d which be obtained from c = cumsum(r) d = cumsum(q) then (5) becomes

    =
    n=1

    rn qn
    n=1

    (8) (9)

    (dm+k1 dm1 )

    1 (cm+k1 cm1 )2 Jk

    (10)

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy