Return to hgm.mleFBByOptim.Rd CVS log

Up to [local] / OpenXM / src / R / r-packages / hgm / man

Annotation of OpenXM/src/R/r-packages/hgm/man/hgm.mleFBByOptim.Rd, Revision 1.3

1.3     ! takayama    1: % $OpenXM: OpenXM/src/R/r-packages/hgm/man/hgm.mleFBByOptim.Rd,v 1.2 2014/03/31 07:23:09 takayama Exp $
        !             2: \name{hgm.z.mleFBByOptim}
        !             3: %%Todo, write documents for hgm.z.mleDemo, hgm.ssFB
1.1       takayama    4: %\alias{hgm.ncso3}
                      5: %- Also NEED an '\alias' for EACH other topic documented here.
                      6: \title{
                      7:    MLE of Fisher-Bingham distribution by optim and HGM.
                      8: }
                      9: \description{
                     10:   It makes the maximal likelihood estimate (MLE) for the Fisher-Bingham
                     11:   distribution on S^d.
                     12: }
                     13: \usage{
1.3     ! takayama   14:  hgm.z.mleFBByOptim(d=0,ss=NULL,data=NULL,start=NULL,lb=NULL,up=NULL,bigValue=10000)
1.1       takayama   15: }
                     16: %- maybe also 'usage' for other objects documented here.
                     17: \arguments{
                     18:   \item{d}{The dimension of the sphere}
                     19:   \item{ss}{Sufficient statistics}
                     20:   \item{data}{
                     21:      The argument data is a set of data on the d-dimensional sphere.
1.2       takayama   22:      Its format is an n by (d+1) matrix where n is the number of data.
                     23:      When data is given, ss must be NULL
1.1       takayama   24:     and ss is calculated from data by hgm.ssFB(data).
                     25:   }
                     26:   \item{start}{
                     27:      Starting point for the function optim. The default value is a random
                     28:      vector.
                     29:   }
                     30:   \item{lb}{
                     31:      An array of length sslen = (d+1)*(d+2)/2 + (d+1), each of which
                     32:      is the lower bound of the parameter. The default value is -100.
                     33:   }
                     34:   \item{ub}{
                     35:      An array of length sslen = (d+1)*(d+2)/2 + (d+1), each of which
                     36:      is the upper bound of the parameter. The default value is 100.
                     37:   }
                     38:   \item{bigValue}{
                     39:      It is used as a value wall to avoid that the evaluation point is out of
                     40:      the search domain defined by lb and ub.
                     41:   }
                     42: }
                     43: \details{
                     44:    It solves the MLE for the Fisher-Bingham distribution.
                     45:    The normalizing constant is evaluated by hgm_ko_ncfb (external program,
                     46:    which should in the path).
                     47:    The function
                     48: %  \code{\link[RCurl]{postForm}}
                     49:   \code{\link{optim}}
                     50:    is used for the optimization.
                     51:    The output is used as a starting point for the holonomic gradient method.
                     52:    See nk_fb_gen_c.rr of \url{http://www.math.kobe-u.ac.jp/Asir}.
1.2       takayama   53:    This function generates temporary work files whose names start with tmp.
                     54:    \code{data <- read.table(fileName,sep=",")} can be used to read CSV data
                     55:    from a file.
1.1       takayama   56: }
                     57: \value{
                     58: The output format is that of the function optim().
                     59: }
                     60: \references{
                     61: T. Koyama, H. Nakayama, K. Nishiyama, N. Takayama,
                     62: Holonomic Gradient Descent for the Fisher-Bingham Distribution on the d-dimensional Sphere,
                     63: Computational Statistics (2013)
                     64: \url{http://dx.doi.org/10.1007/s00180-013-0456-z}
                     65: }
                     66: \author{
                     67: T.Koyama, H.Nakayama, K.Nishiyama, N.Takayama.
                     68: }
                     69: \note{
                     70: %%  ~~further notes~~
                     71: }
                     72:
                     73: %% ~Make other sections like Warning with \section{Warning }{....} ~
                     74:
                     75: \seealso{
1.2       takayama   76: \code{\link{optim}}
1.1       takayama   77: }
                     78: \examples{
                     79: ## =====================================================
                     80: ## Example 1. Asteroid data in [N3OST2]
                     81: ## =====================================================
                     82: \dontrun{
                     83:   d <- 2
                     84:   ss <- c(0.3119,0.0292,0.0707,
                     85:                  0.3605,0.0462,
                     86:                            0.3276,
                     87:             0.0063,0.0054,0.0762)
                     88:   start <- c(0.1,0.1,1,1,1,-1,-1,-1,-1)
1.3     ! takayama   89:   hgm.z.mleFBByOptim(d=d,ss=ss,start=start)
1.1       takayama   90: }
                     91: }
                     92: % Add one or more standard keywords, see file 'KEYWORDS' in the
                     93: % R documentation directory.
                     94: \keyword{ Holonomic gradient method }
                     95: \keyword{ HGM }
                     96: \keyword{ Fisher-Bingham distribution on S^d}
                     97: \keyword{ MLE }
                     98: