waverify2d(SpatialVx)
waverify2d()所属R语言包：SpatialVx
High-Resolution Gridded Forecast Verification Using Discrete Wavelet Decomposition
基于离散小波分解的高分辨率网格化预报检验
译者：生物统计家园网 机器人LoveR
描述----------Description----------
High-resolution gridded forecast verification using discrete wavelet decomposition.
离散小波分解的高分辨率格点预报检验。
用法----------Usage----------
waverify2d(X, Y, Clim = NULL, wavelet.type = "haar", J = NULL, useLL = FALSE, compute.shannon = FALSE, which.space = "field", verbose = FALSE)
mowaverify2d(X, Y, Clim = NULL, wavelet.type = "haar", J = 4, useLL = FALSE, compute.shannon = FALSE, which.space = "field", verbose = FALSE)
## S3 method for class 'waverify2d'
plot(x, main1 = "X", main2 = "Y", main3 = "Climate", which.plots = c("all", "dwt2d", "details", "energy", "mse", "rmse", "acc"), separate = FALSE, ...)
参数----------Arguments----------
参数：X,Y,Clim
m X n dyadic matrices (i.e., m = 2^M and n = 2^N, for M, N some integers) giving the verification and forecast fields (and optionally a climatology field), resp.
m×n的二元矩阵（即m = 2 ^ m和n = 2 ^ N为M，N一些整数）给予验证和预报场（和可选的气候场）分别。
参数：x
list object of class "waverify2d" as returned by waverify2d.
列表对象类“waverify2d”返回waverify2d。
参数：wavelet.type
character naming the type of wavelet to be used. This is given as teh wf argument to the dwt.2d function of package waveslim.
字符命名要使用的类型的小波。这是德wf参数的dwt.2d功能包waveslim的。
参数：J
(optional) numeric integer giving the pre-determined number of levels to use. If NULL, J is set to be log2(m) = M in waverify2d only.
（可选）数字整数的水平，使用预先确定的。如果为NULL，J被设定为在waverify2d只的log2（米）= M。
参数：useLL
logical, should the LL submatrix (i.e., the father wavelet or grand mean) be used to find the inverse DWT's for calculating the detail fields?
逻辑，应LL子矩阵（即小波的父亲或盛大的意思）的逆DWT的计算明细字段？
参数：compute.shannon
logical, should the Shannon entropy be calculated for the wavelet decomposition?
逻辑，香农熵计算的小波分解？
参数：which.space
character (one of "field" or "wavelet") naming from which space the detail fields should be used. If "field", then it is in the original field (or image) space (i.e., the detail reconstruction), and if "wavelet", it will be done in the wavelet space (i.e., the detail wavelet coefficients).
字符（一个“场”或“小波”）的命名应使用其空间的详细信息字段。如果“场”，那么它是在原来的字段（或图像）的空间（即，重建的细节），如果是“小波”，它将会在小波空间（即，详细的小波系数）进行。
参数：main1,main2,main3
optional characters naming each field to be used for the detail field plots and legend labelling on the energy plot.
可选字符命名每个字段的详细信息领域的图和传说中的能量图标签使用。
参数：which.plots
character vector describing which
字符向量，它描述了
参数：separate
logical, should the plots be on their own devices (TRUE) or should some of them be put onto a single multi-panel device (FALSE, default)?
逻辑的，应该图上自己的设备（TRUE），或其中一些被放到一个单一的多面板的移动设备（FALSE，默认值）？
参数：verbose
logical, should progress information be printed to the screen, including total run time?
逻辑的发展，应以信息打印到屏幕上，包括总运行时间？
参数：...
optional additonal plot or image.plot parameters. If detail and energy, mse, rmse or acc plots are desired, must be applicable to both types of plots.
可选的的产生额外的图或image.plot参数。如果细节和能量的均方根误差（MSE），或按图的需要，必须适用于这两种类型的图。
Details
详细信息----------Details----------
This is a function to use discrete wavelet decomposition to analyze verification sets along the lines of Briggs and Levine (1997), as well as Casati et al. (2004) and Casati (2009). In the originally proposed formulation of Briggs and Levine (1997), continuous verification statistics (namely, the anomaly correlation coefficient (ACC) and root mean square error (RMSE)) are calculated for detail fields obtained from wavelet decompositions of each of a forecast and verification field (and for ACC a climatology field as well). Casati et al. (2004) introduced an intensity scale approach that applies 2-d DWT to binary (obtained from thresholding) difference fields (Forecast - Verification), and applying a skill score at each level based on the mean square error (MSE). Casati (2009) extended this idea to look at the energy at each level as well.
这是一个函数，使用离散小波分解分析验证集布里格斯和Levine（1997），以及CASATI等沿线。 （2004年）和CASATI（2009年）。在最初提出的制定布里格斯和Levine（1997年），连续检验统计（即距平相关系数（ACC）和均方根误差（RMSE））的详细信息字段从小波分解的预测和计算，验证字段（和一个气候学领域，以及ACC）。卡萨蒂等。 （2004）介绍了强度等级为二进制的方法，适用于2-D DWT（阈值）获得差异字段（预测 - 验证），并在每个级别均方误差（MSE）的基础上应用技巧得分。 CASATI（2009）扩展了这一想法，看在每个级别的能源。
This function makes use of the dwt.2d and idwt.2d functions from package waveslim, and plot.waverify2d uses the plot.dwt.2d function if dwt2d is selected through the which.plots argument. See the help file for these functions, the references therein and the references herein for more on these approaches.
这函数使用的从包装waveslim功能dwt.2d和idwt.2d的，和plot.waverify2d使用的plot.dwt.2d的功能，如果dwt2d选择通过which.plots参数。这些功能，参考文献，文中提及这些方法请参阅帮助文件。
Generally, it is not necessary to use the father wavelet for the detail fields, but for some purposes, it may be desired.
一般来说，它是没有必要使用的父亲小波的细节信息，但为了某些目的，可能希望。
mowaverify2d is very similar to waverify2d, but it allows fields to be non-dyadic (and may subsequently be slower). It uses the modwt.2d and imodwt.2d functions from the package waveslim. In particular, it performs a maximal overlap discrete wavelet transform on a matrix of arbitrary dimension. See the help file and references therein for modwt.2d for more information, as well as Percival and Guttorp (1994) and Lindsay et al. (1996).
mowaverify2d是非常相似waverify2d的，但它允许字段非二进（及随后可能会比较慢）。它采用了modwt.2d和imodwt.2d的函数从包waveslim。特别是，它执行一个最大重叠离散小波变换的任意维数的矩阵上。请参阅帮助文件和参考其中为modwt.2d的更多信息，以及作为波斯富街及Guttorp的（1994）和林赛等。 （1996年）。
In Briggs and Levine (1997), they state that the calculations can be done in either the data (called field here) space or the wavelet space, and they do their examples in the field space. If the wavelets are orthogonal, then the detail coefficeints (wavelet space), can be analyzed with the assumption that they are independent; whereas in the data space, they typically cannot be assumed to be independent. Therefore, most statistical tests should be performed in the wavelet space to avoid issues arising from spatial dependence.
在布里格斯和Levine（1997），他们的状态计算可以做到无论是在空间或小波空间数据（这里称为字段），而在场地空间，他们做他们的榜样。如果小波是正交的，那么的的详细coefficeints（小波空间），可以分析与假设它们是独立的，而在数据空间中，它们通常不能被假定为独立。因此，大多数统计测试应在小波空间，以避免问题产生的空间依赖性。
值----------Value----------
A list object of class "waverify2d" with components:
一个List对象类“waverify2d”的组件：
参数：J
single numeric giving the number of levels.
单数字的数量级。
参数：X.wave, Y.wave, Clim.wave
objects of class "dwt.2d" describing the wavelet decompositions for the verification and forecast fields (and climatology, if applicable), resp. (see the help file for dwt.2d from package waveslim for more about these objects).
“dwt.2d类的对象进行验证和预测领域和气候（如适用）分别描述了小波分解。 （见的包waveslim的帮助文件dwt.2d从这些对象）。
参数：Shannon.entropy
numeric matrix giving the Shannon entropy for each field.
数字矩阵的每个字段的香农熵。
参数：energy
numeric matrix giving the energy at each level and field.
数字矩阵的能量在各层次和领域。
参数：mse,rmse
numeric vectors of length J giving the MSE/RMSE for each level between the verification and forecast fields.
数字向量的长度为J的MSE / RMSE为每个级别之间的验证和预测等领域。
参数：acc
If a climatology field is supplied, this is a numeric vector giving the ACC for each level.
如果气候字段被供给的，这是一个给ACC为每个级别的数值向量。
（作者）----------Author(s)----------
Eric Gilleland
参考文献----------References----------
参见----------See Also----------
dwt.2d, idwt.2d, hoods2d, hoods2dPrep
dwt.2d，idwt.2d，hoods2d，hoods2dPrep
实例----------Examples----------
data(UKobs6)
data(UKfcst6)
look <- waverify2d(UKobs6, UKfcst6)
plot(look, which.plots="energy")
look2 <- mowaverify2d(UKobs6, UKfcst6, J=8)
plot(look2, which.plots="energy")
## Not run: [＃不运行：]
pdf("dyadicDWTex.pdf")
plot(look, main1="NIMROD Analysis", main2="NIMROD Forecast")
dev.off()
pdf("nondyadicMODWTex.pdf")
plot(look2, main1="NIMROD Analysis", main2="NIMROD Forecast")
dev.off()
data(pert000)
data(pert004)
look <- mowaverify(pert000, pert004, J=8, verbose=TRUE) # Slow, but does not require fields to be dyadic.[慢，但并不要求字段是二进。]
plot(look, which.plots="energy") # Also can just do plot(look), but should print to a pdf file (e.g., using pdf()).[也只是做图（外观），但应打印到PDF文件（例如，使用PDF（））。]
# Try one with some kind of climatology field. Here using surrogater2d function.[请尝试某种气候领域之一。这里使用surrogater2d功能。]
data(UKloc)
hold <- surrogater2d(UKobs6, n=1, maxiter=50, verbose=TRUE)
hold <- matrix(hold, 256, 256)
image(hold, col=c("grey",tim.colors(64)), axes=FALSE)
image.plot(UKloc, col=c("grey",tim.colors(64)), legend.only=TRUE, horizontal=TRUE)
look <- waverify2d(UKobs6, UKfcst6, hold)
pdf("waveletEx.pdf")
plot(look)
dev.off()
## End(Not run)[＃（不执行）]
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 2012-9-30 12:59:28
