SDisc(SDisc)
SDisc()所属R语言包：SDisc
SDisc to discover homogeneous subtypes in data
SDisc发现同质亚型，数据
译者：生物统计家园网 机器人LoveR
描述----------Description----------
Performs and returns an SDisc analysis on the data. This analysis involves repeated mixture modeling for different combinations of number of components, number of mixture model parameters and random initialization start. The SDisc results contains a data set container (SDData) which stores the original data and which may, e.g., limit the cluster analysis to a few variables (see SDData). It contains, too, the parameters of the different models estimated and a BIC table summarizing their likelihood and rank. Generic plot, print and summary function enable to visualize and summarize the results (see plot.SDisc, print.SDisc and summary.SDisc).
执行和返回的SDisc的data分析。此分析涉及重复的混合模型的元件数量的不同的组合，混合模型参数的数量和随机的初始化开始。 SDisc结果中包含一个数据集容器（SDData）中存储原始数据和可能，例如，限制到几个变量聚类分析（见SDData“）。据载，太，参数估计的不同型号和BIC的可能性和排名表总结的。通用的图，打印和汇总功能，使可视化和总结的结果（见plot.SDisc，print.SDisc和summary.SDisc）。
用法----------Usage----------
## Default S3 method:[默认方法]
SDisc(x, cfun="modelBasedEM", cFunSettings=list(modelName = c("EII", "VII"), G = 3:5, rseed = 6013:6015), nTopModels=5, nnodes=1, ...)
## S3 method for class 'SDisc'
plot(x, q=NULL, type=c('plotParcoord', 'plotLegend', 'plotPC1', 'plotPC2', 'plotDendroCluster',
'plotDendroVar'), latex=FALSE, title=NULL, xlim=c(-3, 3), zlim=c(-2, 2), xy=c(-2.2, 0), pattern=mean, cex=0.7,
colGrad=rev(brewer.pal(9, "RdBu")), rangeFV=NULL, lab=NULL, ...)
## S3 method for class 'SDisc'
predict(object, newdata, ...)
## S3 method for class 'SDisc'
print(x, y=NULL, m1=1, m2=2, latex=FALSE, lab="jointdistrib", ...)
## S3 method for class 'SDisc'
summary(object, q = 1, ...)
参数----------Arguments----------
参数：x
a data matrix (with its settings file), an SDData instance or an SDisc data object
数据矩阵（它的settings文件），SDData实例或SDisc数据对象
参数：cfun
the name of the cluster algorithm
聚类算法的名称
参数：cFunSettings
the set of parameters of the cluster algorithm
在聚类算法的参数的集合
参数：nTopModels
the number of top-ranking models
数排名靠前的车型
参数：nnodes
the number of nodes in the case of parallel computing
在并行计算的情况下，节点的数目
参数：object
an SDisc analysis result
SDisc的的分析结果
参数：q
an numeric value telling how many top ranking models to characterize graphically (bestModel, a character vector referring to the names of an SDCModel, by default set to NULL that is, the 5 most likely models)
的数值，告诉多少一流的模型来描述图形（bestModel，字符向量的SDCModel，默认设置为NULL，最有可能的模型）的名字
参数：y
an optional second SDisc data object whose model where estimated on the same data
可选的第二个SDisc数据对象的模型估计相同的数据
参数：m1
rank (integer) or name (character vector) of the first model to compare. The rank is passed to bestModel to retrieve the appropriate model name.
排名（整数）或名称（字符向量）的第一款车型比较。排名传递给bestModel来检索相应的模型名称。
参数：m2
rank (integer) or name (character vector) of the second model to compare. The rank is passed to bestModel to retrieve the appropriate model name.
排名（整数）或名称（字符向量）的第二个模型比较。排名传递给bestModel来检索相应的模型名称。
参数：latex
either TRUE or FALSE, whether the LaTeX code must be reported on the standard output for dynamic report generation (Sweave)
TRUE或FALSE，无论是在标准输出上生成动态报告（SweaveLaTeX的代码必须报）
参数：lab
the label of that table
该表的标签
参数：type
a character vector in 'plotParcoord', 'plotLegend', 'plotImage', 'plotDendroCluster', 'plotDendroVar'
一个字符向量'plotParcoord', 'plotLegend', 'plotImage', 'plotDendroCluster', 'plotDendroVar'
参数：title
the title of the graphics
图形的标题
参数：xlim
the x-limits of the parallel coordinate plots
的x-限制的平行坐标图
参数：zlim
the z-limits for the color gradient in the image
在图像中的颜色渐变的z-限制
参数：xy
the xy-location of the legend
XY位置的传说
参数：pattern
the name of the function to calculate the characteristic pattern, by default mean
的函数名来计算的特征模式，默认情况下，mean
参数：cex
the character expansion numeric value, by default 0.7 70%
字符扩展的数值，默认情况下，0.7 70％
参数：colGrad
a character vector of the colors to use in the color image
在彩色图像的颜色使用一个字符矢量
参数：rangeFV
the range of features when plotting series of values
策划一系列值范围内的功能时，
参数：newdata
an SDData or SDisc object
SDData或SDisc对象
参数：...
additional parameters passed to the SDData function when a data matrix is provided. Such parameters may be settings and prefix.
额外的参数传递给SDData功能，当一个data矩阵。这些参数可能是settings和prefix。
Details
详细信息----------Details----------
plot characterizes graphically the best model (q=NULL) of an SDisc analysis through parallel coordinate, legend, image, sample- and variable-dendrograms. Some plots like heatmap and parallel coordinates admit parameterization, which can be specified in the data description (see SDDataSettings(x)). Generated graphics are in PDF and Sweave output can be switched on with latex=TRUE. Additional parameters influence the color scales, range of values.
plot图形化的最佳模式（q=NULL）的SDisc分析，通过平行坐标，传说，图片，样品和可变树状图的特点。一些图，如热图和平行坐标承认的参数，可以指定数据描述（见SDDataSettings(x)）。生成图形在PDF和Sweave输出可以切换上与latex=TRUE。其他参数影响的色阶，数值范围。
predict estimates a set of subtypes (mixture models) on new data based on the most likely mixture model from an SDisc analysis. Used in combination with print, the new data can help validate a previous model of subtypes.
predict估计最有可能从SDisc分析的混合模型对新数据的基础上的一组子（混合模式）。与print结合使用，新的数据可以帮助验证以前的型号亚型。
print reports the joint distribution between pairs of subtyping results, the agreement (kappa, V chi2-based measure, random index). It is also possible to compare affectation between SDisc analyses with different parameters but with the same data container. In which case, it illustrates how, under different settings, the subtyping analyses agree.
print报告之间的联合分布对分型结果，该协议（KAPPA，Vχ^ 2为基础的措施，随机指标）。另外，也可以使用不同的参数，但具有相同的数据容器之间比较做作SDisc分析。在这种情况下，它说明了如何，根据不同的设置，分型分析的同意。
summary lists the dataset characteristics, the top-ranked models, the BIC table, and oddratios statistics for the top ranked models.
summary列出了数据集的特点，世界排名第一的车型，BIC表，和oddratios统计，排名靠前的车型。
（作者）----------Author(s)----------
Fabrice Colas
参考文献----------References----------
Marinus, H, Cannon, EO, Bender, A, van Hilten, JJ, Slagboom, PE, Kok, JN: A Scenario Implementation in R for Subtype Discovery Examplified on Chemoinformatics Data. Leveraging Applications of Formal Methods, Verification and Validation (ISoLA'08), October 13-15, 2008. [http://dx.doi.org/10.1007/978-3-540-88479-8_48]
Colas, F, Meulenbelt, I, Houwing-Duistermaat, JJ, Kloppenburg, M, Watt, I, van Rooden, SM, Visser, M, Marinus, H, van Hilten, JJ, Slagboom, PE, Kok, JN: Stability of Clusters for Different Time Adjustments in Complex Disease. Research 30th Annual International IEEE EMBS Conference (EMBC'08), Vancouver, Canada, 2008. [http://dx.doi.org/10.1109/IEMBS.2008.4650238]
Fraley C, Raftery AE: Model-Based Clustering, Discriminant Analysis and Density Estimation. Journal of the American Statistical Association, vol. 97, pp. 611-631, 2002. [http://www.stat.washington.edu/raftery/Research/PDF/fraley2002.pdf]
Fraley C, Raftery AE, MCLUST Version 3 for R: Normal Mixture Modeling and Model-Based Clustering. Technical Report 504, Department of Statistics, University of Washington, September 2006. [http://www.stat.washington.edu/fraley/mclust/tr504.pdf]
参见----------See Also----------
SDStability, SDData, SDDataSettings, SDisc,
SDStability，SDData，SDDataSettings，SDisc，
实例----------Examples----------
settings <- SDDataSettings(iris)
settings['Species',] <- c(NA,FALSE, NA, NA, NA, NA)
x <- SDisc(iris, settings=settings, prefix='iris')
### do not run[＃＃不运行]
#plot(x)[图（X）]
print(x)
summary(x)
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 2012-9-29 23:33:07
