AlignNetworks(GraphAlignment)
AlignNetworks()所属R语言包：GraphAlignment
Align networks
对齐网络
译者：生物统计家园网 机器人LoveR
描述----------Description----------
Align networks A and B.
对准网络A和B。
用法----------Usage----------
AlignNetworks(A, B, R, P, linkScore, selfLinkScore, nodeScore1,
nodeScore0, lookupLink, lookupNode, bStart, bEnd, maxNumSteps, clamp=TRUE,
directed=FALSE)
参数----------Arguments----------
参数：A
adjacency matrix for network A
邻接矩阵为网络A
参数：B
adjacency matrix for network B
网络B的邻接矩阵
参数：R
node similarity matrix
节点相似矩阵
参数：P
permutation vector to be used as the initial alignment (see InitialAlignment)
置换向量被用来作为初始对准（见InitialAlignment）
参数：linkScore
link score matrix (see ComputeLinkParameters)
链接得分矩阵（见ComputeLinkParameters）
参数：selfLinkScore
self link score matrix (see ComputeLinkParameters)
自链接得分矩阵（见ComputeLinkParameters）
参数：nodeScore1
node score vector (s1) (see ComputeNodeParameters)
节点的得分矢量（S1）（见ComputeNodeParameters）
参数：nodeScore0
node score vector for unaligned nodes (s0) (see ComputeNodeParameters)
未对齐节点的节点得分向量（S0）（见ComputeNodeParameters）
参数：lookupLink
link bin lookup table (see GetBinNumber)
链接斌查找表（见GetBinNumber）
参数：lookupNode
node bin lookup table (see GetBinNumber)
节点槽查找表（见GetBinNumber）
参数：bStart
start scaling value for simulated annealing
开始扩大为模拟值退火
参数：bEnd
end scaling value for simulated annealing
结束模拟缩放值退火
参数：maxNumSteps
maximum number of steps
最大步数
参数：clamp
clamp values to range when performing bin lookups
钳位值范围时执行的bin查找
参数：directed
whether input networks should be treated as directed graphs
输入网络是否应被视为有向图
Details
详情----------Details----------
This function finds an alignment between the two input networks, specified in the form of adjacency matrices, by repeatedly calling ComputeM and LinearAssignment, up to maxNumSteps times. Simulated annealing is performed if a range is specified in the bStart and bEnd arguments. This simple procedure is described in detail in [Berg, Laessig 2006]. Different procedures can easily be implemented by the user.
这个功能，发现两个输入网络之间的对齐，在邻接矩阵的形式指定，一边喊ComputeM和LinearAssignment，最多maxNumSteps倍。如果在bStart和弯曲参数指定的范围内进行模拟退火。 [伯格，2006年Laessig]中详细描述了这个简单的程序。不同的程序可以很容易地实现用户。
In each step, the matrix M is calculated from the scoring parameters and the current permutation vector P. The result is then normalized to the range [-1, 1] and, if simulated annealing is enabled, a random matrix depending on the current simulated annealing parameters is added. The linear assignment routine is used to calculate the value of P which is used to compute M in the next step.
矩阵M中的每一步，从得分参数计算，目前置换向量体育结果，然后标准化的范围为[-1，1]，如果启用模拟退火，随机矩阵根据对当前的模拟退火参数被添加。线性分配程序是用来计算P值，这是用来计算在下一步中号。
If the flag directed is set, directed binary networks are encoded by suitable symmetric matrices using EncodeDirectedGraph. The corresponding 3x3 matrices of the link score are computed from the 2x2 matrices given as input.
如果指示标志设置，指示的二进制网络编码使用EncodeDirectedGraph适合对称矩阵。计算相应的链接得分的3x3矩阵作为输入给定的2x2矩阵。
Simulated annealing is enabled if bStart differs from bEnd. In this case, a value bStep = bEnd - bStart) / (maxNumSteps - 1) is calculated. In step n, the random matrix which is added to M is scaled by the factor 1 / [bStart + (n - 1) * bStep].
模拟退火启用如果bStart不同弯曲。在这种情况下，值bStep弯 - bStart）/（maxNumSteps - 1）计算。在第n步，添加到随机矩阵M是缩放因子1 / [bStart +（N - 1）* bStep]。
值----------Value----------
The return value is a permutation vector p which aligns nodes from network a with nodes from network B (including dummy nodes). The returned permutation should be read in the following way: the node i in the network A is aligned to that node in the network B which label is at the i-th position of the permutation vector p. If the label at this position is larger than the size of the network B, the node i is not aligned.
返回值是一个置换向量p赞同从网络，从网络B（包括假人节点）的节点1节点。返回置换应读通过以下方式：在网络节点我一个置换向量p i个位置的标签，在网络B的节点对齐。如果在这个位置上的标签是比网络B的大小，节点i没有对齐。
作者（S）----------Author(s)----------
Joern P. Meier, Michal Kolar, Ville Mustonen, Michael Laessig, and Johannes Berg
参考文献----------References----------
举例----------Examples----------
ex<-GenerateExample(dimA=22, dimB=22, filling=.5, covariance=.6,
symmetric=TRUE, numOrths=10, correlated=seq(1,18))
pinitial<-InitialAlignment(psize=34, r=ex$r, mode="reciprocal")
lookupLink<-seq(-2,2,.5)
linkParams<-ComputeLinkParameters(ex$a, ex$b, pinitial, lookupLink)
lookupNode<-c(-.5,.5,1.5)
nodeParams<-ComputeNodeParameters(dimA=22, dimB=22, ex$r,
pinitial, lookupNode)
al<-AlignNetworks(A=ex$a, B=ex$b, R=ex$r, P=pinitial,
linkScore=linkParams$ls,
selfLinkScore=linkParams$ls,
nodeScore1=nodeParams$s1, nodeScore0=nodeParams$s0,
lookupLink=lookupLink, lookupNode=lookupNode,
bStart=.1, bEnd=30,
maxNumSteps=50)
转载请注明:出自 生物统计家园网(http://www.biostatistic.net)。
注：
注1：为了方便大家学习，本文档为生物统计家园网机器人LoveR翻译而成，仅供个人R语言学习参考使用，生物统计家园保留版权。
注2：由于是机器人自动翻译，难免有不准确之处，使用时仔细对照中、英文内容进行反复理解，可以帮助R语言的学习。
注3：如遇到不准确之处，请在本贴的后面进行回帖，我们会逐渐进行修订。
|
发表于 2012-2-25 20:58:31
