[Gsdesign-commits] r162 - pkg/tex

[noreply at r-forge.r-project.org]

Author: keaven
Date: 2009-05-22 23:32:33 +0200 (Fri, 22 May 2009)
New Revision: 162

Removed:
   pkg/tex/gsDesignManual.toc
   pkg/tex/install.tex
   pkg/tex/objtypes.tex
   pkg/tex/other_doc.tex
   pkg/tex/output.tex
Log:
v 2 manual completion

Deleted: pkg/tex/gsDesignManual.toc
===================================================================
--- pkg/tex/gsDesignManual.toc	2009-05-22 21:32:07 UTC (rev 161)
+++ pkg/tex/gsDesignManual.toc	2009-05-22 21:32:33 UTC (rev 162)
@@ -1,49 +0,0 @@
-\contentsline {section}{\numberline {1}Overview}{2}{section.1}
-\contentsline {section}{\numberline {2}Basic Features}{3}{section.2}
-\contentsline {section}{\numberline {3}Installation and Online Help}{4}{section.3}
-\contentsline {section}{\numberline {4}Detailed Examples}{4}{section.4}
-\contentsline {section}{\numberline {5}Syntax}{17}{section.5}
-\contentsline {subsection}{\numberline {5.1}gsDesign() syntax}{18}{subsection.5.1}
-\contentsline {subsection}{\numberline {5.2}gsProbability() syntax}{21}{subsection.5.2}
-\contentsline {subsection}{\numberline {5.3}Conditional power:\ gsCP() and gsBoundCP() syntax}{21}{subsection.5.3}
-\contentsline {subsection}{\numberline {5.4}Binomial distribution: FarrMannSS(), MandNtest(), MandNsim() syntax}{22}{subsection.5.4}
-\contentsline {section}{\numberline {6}The gsDesign and gsProbability Object Types}{24}{section.6}
-\contentsline {subsection}{\numberline {6.1}The spendfn Class}{24}{subsection.6.1}
-\contentsline {subsection}{\numberline {6.2}The gsProbability Class}{25}{subsection.6.2}
-\contentsline {subsection}{\numberline {6.3}The gsDesign Class}{25}{subsection.6.3}
-\contentsline {section}{\numberline {7}Formatted Output}{26}{section.7}
-\contentsline {section}{\numberline {8}Spending Functions}{27}{section.8}
-\contentsline {subsection}{\numberline {8.1}Spending Function Basics}{27}{subsection.8.1}
-\contentsline {subsection}{\numberline {8.2}Spending Function Details}{28}{subsection.8.2}
-\contentsline {subsection}{\numberline {8.3}Investigational Spending Functions}{30}{subsection.8.3}
-\contentsline {subsection}{\numberline {8.4}Resetting timing of analyses}{32}{subsection.8.4}
-\contentsline {subsection}{\numberline {8.5}Optimized Spending Functions}{33}{subsection.8.5}
-\contentsline {subsection}{\numberline {8.6}Writing Code for a New Spending Function}{35}{subsection.8.6}
-\contentsline {section}{\numberline {9}Statistical Methods}{36}{section.9}
-\contentsline {section}{\numberline {10}Function and Class Reference}{41}{section.10}
-\contentsline {subsection}{\numberline {10.1}gsDesign Package}{41}{subsection.10.1}
-\contentsline {subsection}{gsDesign-package}{41}{subsection.10.1}
-\contentsline {subsection}{\numberline {10.2}gsDesign main functions}{42}{subsection.10.2}
-\contentsline {subsection}{gsDesign}{42}{subsection.10.2}
-\contentsline {subsection}{gsProbability}{46}{section*.23}
-\contentsline {subsection}{plot.gsDesign}{48}{section*.33}
-\contentsline {subsection}{gsCP}{51}{section*.43}
-\contentsline {subsection}{gsBoundCP}{53}{section*.53}
-\contentsline {subsection}{\numberline {10.3}Binomial trial functions}{54}{subsection.10.3}
-\contentsline {subsection}{normalGrid}{54}{subsection.10.3}
-\contentsline {subsection}{nBinomial}{56}{section*.72}
-\contentsline {subsection}{nSurvival}{60}{section*.80}
-\contentsline {subsection}{\numberline {10.4}Spending Functions}{62}{subsection.10.4}
-\contentsline {subsection}{Spending Functions}{62}{subsection.10.4}
-\contentsline {subsection}{sfHSD}{65}{section*.98}
-\contentsline {subsection}{sfPower}{66}{section*.108}
-\contentsline {subsection}{sfExponential}{68}{section*.118}
-\contentsline {subsection}{sfLDOF}{70}{section*.128}
-\contentsline {subsection}{sfPoints}{72}{section*.138}
-\contentsline {subsection}{sfLogistic}{73}{section*.147}
-\contentsline {subsection}{sfTDist}{76}{section*.157}
-\contentsline {subsection}{\numberline {10.5}Other Files}{78}{subsection.10.5}
-\contentsline {subsection}{Wang-Tsiatis Bounds}{78}{subsection.10.5}
-\contentsline {subsection}{checkScalar}{79}{section*.174}
-\contentsline {subsection}{gsbound}{81}{section*.179}
-\contentsline {section}{\numberline {11}Contacts}{84}{section.11}

Deleted: pkg/tex/install.tex
===================================================================
--- pkg/tex/install.tex	2009-05-22 21:32:07 UTC (rev 161)
+++ pkg/tex/install.tex	2009-05-22 21:32:33 UTC (rev 162)
@@ -1,39 +0,0 @@
-\section{Installation and Online Help\label{sec:install}}
-
-The package comes in a binary format for a Windows platform in the file
-gsDesign-2.0.zip (may be updated to fix bugs in a file such as
-gsDesign-2.0.zip). This file includes a copy of this manual in the file
-gsDesignManual.pdf. The source, available for other platforms (but only tested
-minimally!) is in the file gsDesign-2.0.tar.gz (may be updated to fix bugs in
-a file such as gsDesign-2.0.tar.gz). Following are basic instructions for
-installing the binary version on a Windows machine. It is assumed that a
-`recent' version of R is installed. From the Windows interface of R, select
-the Packages menu line and from this menu select Install packages from local
-zip files\ldots. Browse to select gsDesign-2.0.zip. Once installed, you need
-to load the package by selecting the Packages menu line, selecting Load
-package\ldots\ from this menu, and then selecting gsDesign. You are now ready
-to use the routines in the package. The most up-to-date version of this manual
-and the code is also available at \texttt{http://r-forge.r-project.org}.
-
-\bigskip
-
-Online help can be obtained by entering the following on the command line:
-
-\bigskip
-
-\texttt{%
-%TCIMACRO{\TEXTsymbol{>} }%
-%BeginExpansion
-$>$
-%EndExpansion
-help(gsDesign)}
-
-\bigskip
-
-There are many help topics covered there which should be sufficient
-information to keep you from needing to use this document for day-to-day use.
-In the Window version of R, this brings up a "Contents" window and a
-documentation window. In the contents window, open the branch of documentation
-headed by "Package gsDesign: Titles." The help files are organized
-sequentially under this heading. 
-

Deleted: pkg/tex/objtypes.tex
===================================================================
--- pkg/tex/objtypes.tex	2009-05-22 21:32:07 UTC (rev 161)
+++ pkg/tex/objtypes.tex	2009-05-22 21:32:33 UTC (rev 162)
@@ -1,162 +0,0 @@
-\section{The gsDesign and gsProbability Object Types\label{sec:objtypes}}
-
-This package creates three object (storage) classes to store the structured
-information returned by \texttt{gsDesign()}, \texttt{gsProbability()} and
-spending function routines. Thus, a single variable name can be used to
-access, print or plot a set of returned values. \bigskip
-
-\subsection{The spendfn Class\label{sec:spendfn}}
-
-\bigskip
-
-The first class, \texttt{spendfn}, is used to store information about either 
-the upper or the lower bound of a design: the spending function, z-values for 
-the stopping bounds, spending at each analysis and boundary crossing 
-probabilities under parameter values of interest. The \texttt{spendfn} class
-contains the following members:
-
-\begin{itemize}
-\item \texttt{sf}: a spending function or character value. Normally this
-is a spending function, but if Wang-Tsiatis ("WT"), Pocock ("Pocock"), or
-O'Brien-Fleming ("OF") are specified (these are not spending function
-designs), then \texttt{sf} is a character string. This is useful if you
-later wish to plot the spending function for a larger set of values than
-specified in an original call. ({\bf Note for writers of new spending 
-functions:} this is not set in the spending function itself, but in 
-\texttt{gsDesign()})
-
-\item \texttt{name}: a character string specifying the spending function used.
-
-\item \texttt{param}: parameter value or values to fully specify the
-spending function using the spending function family specified in \texttt{sf}.
-
-\item \texttt{parname}: a text string or vector of text strings (matching
-the length of \texttt{param}) supplying the names of parameters.
-
-\item \texttt{spend}: a vector containing the amount of $\alpha$- or $\beta
-$-spending at each analysis; this is determined in a call to
-\texttt{gsDesign()}.
-
-\item \texttt{bound}: this is null when returned from the spending function,
-but is set in \texttt{gsDesign()} if the spending function is called from
-there. Contains z-values for bounds of a design.
-
-\item \texttt{prob}: this is null when returned from the spending function,
-but is set in \texttt{gsDesign()} if the spending function is called from
-there. Contains a \texttt{k}$\times$\texttt{n} matrix of probabilities of
-boundary crossing at\texttt{\ i}-th analysis for \texttt{j}-th \texttt{theta}
-value in \texttt{prob[i,j]}. Upon return from \texttt{gsDesign()},
-\texttt{n}=2 and the values of theta considered are 0 and \texttt{delta}. More
-values of \texttt{theta} can be added by a call to \texttt{gsProbability()}.
-Columns correspond to the values specified by \texttt{theta}. All boundary
-crossing is assumed to be binding for this computation; that is, the trial
-must stop if a boundary is crossed.
-
-\item \texttt{errcode}: 0 if no error in spending function specification,
-%TCIMACRO{\TEXTsymbol{>} }%
-%BeginExpansion
-$>$
-%EndExpansion
-0 otherwise.
-
-\item \texttt{errmsg}: a text string corresponding to \texttt{errcode}; for
-no error, this is "No error."
-\end{itemize}
-
-\subsection{The gsProbability Class}
-
-\bigskip
-
-The second object class is \texttt{gsProbability}. The members of this class 
-on output from \texttt{gsProbability()} are: \texttt{k}, \texttt{n.I}, 
-\texttt{lower}, \texttt{upper}, \texttt{r}, \texttt{theta}, \texttt{errcode}, 
-and \texttt{errmsg}. Most of these are described in the input to
-\texttt{gsProbability()}. The exceptions are as follows:
-
-\begin{itemize}
-\item \texttt{upper}: on output, contains information on upper spending;
-this is a member of the \texttt{spendfn} class described above. Values of
-\texttt{upper} returning in a \texttt{gsProbability} class contain only the
-elements \texttt{bound} and \texttt{prob}.
-
-\item \texttt{lower}: on output, contains information on lower spending;
-this is a member of the \texttt{spendfn} class described above. Values of
-\texttt{lower} returning in a \texttt{gsProbability} class contain only the
-elements \texttt{bound} and \texttt{prob}.
-
-\item \texttt{en}: a vector of the same length as \texttt{theta} containing
-expected sample sizes for the trial design corresponding to each value in the
-vector \texttt{theta}.
-
-\item \texttt{errcode}: 0 if no error detected in call to
-\texttt{gsDesign()} or \texttt{gsProbability()},
-%TCIMACRO{\TEXTsymbol{>} }%
-%BeginExpansion
-$>$
-%EndExpansion
-0 otherwise.
-
-\item \texttt{errmsg}: a text string corresponding to \texttt{errcode}; for
-no error, this is "No errors detected."
-\end{itemize}
-
-\subsection{The gsDesign Class}
-
-The third and final object class is \texttt{gsDesign}. This inherits the class
-\texttt{gsProbability} and, in addition, has the following members on output from
-\texttt{gsDesign()}: \texttt{test.type}, \texttt{alpha}, \texttt{beta},
-\texttt{delta}, \texttt{n.fix}, \texttt{timing}, \texttt{tol}, 
-\texttt{maxn.IPlan}. In
-addition, all elements of the class 
-\text{spendfn} are returned in \texttt{upper} and
-\texttt{lower}. Most of these variables are as described in the input to
-\texttt{gsDesign()} or in the \texttt{spendfn} or \texttt{gsProbability} 
-class description. The exceptions are as follows:
-
-\begin{itemize}
-\item \texttt{timing}: when this is input as 1, the output is transformed to
-a vector of equally spaced analyses as follows: (1, 2, \ldots, \texttt{k}%
-)/\texttt{k}. Otherwise, this should be a vector of length \texttt{k}, with
-increasing values greater than 0 and the greatest value equal to 1. A value of
-\texttt{timing[2] = 0.4} is translated as the second interim analysis being performed
-after 40\% of the total observations planned for the final analysis.
-
-\item \texttt{delta}: the standardized effect size; if this was input as 0,
-it is recomputed with the following formula:%
-\[
-\delta=\frac{\Phi^{-1}(1-\alpha)+\Phi^{-1}(1-\beta)}{\sqrt{N_{fix}}}
-\]
-
-
-where $\delta$=\texttt{delta}, $N_{fix}$=\texttt{n.fix}, $\alpha
-$=\texttt{alpha}, and $\beta$=\texttt{beta}.
-
-\item \texttt{n.I}: this is the sample size or information required at each
-analysis, depending on how \texttt{gsDesign()} is called. In addition to the
-following descriptions, see Section~\ref{sec:detailedex}, Detailed Examples.
-
-\begin{enumerate}
-\item If \texttt{n.I} was input, the same value remains on output
-
-\item If input to \texttt{gsDesign()} was \texttt{n.fix} = 1 and
-\texttt{delta}=0, this results in returning a sample-size multiplier in
-\texttt{n.I} that can be used to set sample size at interim analyses according
-to the sample size required for a fixed design without interim analysis. That
-is, use a formula for a fixed design that matches the type of data you are
-collecting, and multiply this fixed design sample size by \texttt{n.I} to
-obtain sample sizes at interim and final analyses for the desired group
-sequential design.
-
-\item Similarly, if you have a fixed design sample size which you wish to
-modify using \texttt{gsDesign()} to obtain an appropriate group sequential
-design, input \texttt{n.fix} as the fixed sample size design and input
-\texttt{delta}=0. In this case \texttt{n.I} returns with sample sizes for
-each analysis in the group sequential design.
-\end{enumerate}
-
-\item \texttt{upper}, \texttt{lower}: these are output as class 
-\texttt{spendfn}, as described above. All elements of the 
-\texttt{spendfn} class are returned when these are
-generated from a call to \texttt{gsDesign()}.
-\end{itemize}
-

Deleted: pkg/tex/other_doc.tex
===================================================================
--- pkg/tex/other_doc.tex	2009-05-22 21:32:07 UTC (rev 161)
+++ pkg/tex/other_doc.tex	2009-05-22 21:32:33 UTC (rev 162)
@@ -1,3 +0,0 @@
-\subsection{Other Files}\input{./tmphelp/tex/Wang-Tsiatis-bounds}
-\input{./tmphelp/tex/checkScalar}
-\input{./tmphelp/tex/gsbound}

Deleted: pkg/tex/output.tex
===================================================================
--- pkg/tex/output.tex	2009-05-22 21:32:07 UTC (rev 161)
+++ pkg/tex/output.tex	2009-05-22 21:32:33 UTC (rev 162)
@@ -1,81 +0,0 @@
-\section{Formatted Output\label{sec:formatted}}
-
-\texttt{gsDesign()} returns an object of the class \texttt{gsDesign}.
-\texttt{gsProbability()} returns an object of class gsDesign if a \texttt{gsDesign} class object is passed in or of class \texttt{gsProbability}, if not. 
-The standard R functions \texttt{print()}, and \texttt{plot()} are extended 
-to work for both the \texttt{gsDesign} and the \texttt{gsProbability} classes. 
-Note also that \texttt{summary()} prints a brief summary of either object 
-type if you need a reminder of what is
-in a class. The \texttt{print()} function for each object class has a single
-argument and is implemented through the functions \texttt{print.gsDesign()}
-and \texttt{print.gsProbability()}. The \texttt{plot()} functions for 
-\texttt{gsDesign} and \texttt{gsProbability} objects have a second argument 
-that specifies which type of
-plot is to be made; these functions are implemented with the functions
-\texttt{plot.gsDesign()} and \texttt{plot.gsProbability()}, respectively.
-There are seven plot types, six of which are available for both classes, 
-that are specified through the input variable \texttt{plottype}:
-
-\bigskip
-
-\begin{itemize}
-\item \texttt{1} or \texttt{"Z"}: boundary plots (default if
-\texttt{class(x)="gsDesign"})
-
-\item \texttt{2} or \texttt{"power"}: boundary crossing probability plots
-(default if \texttt{class(x)="gsProbability"})
-
-\item \texttt{3} or \texttt{"thetahat"}: estimated treatment effect at boundaries
-
-\item \texttt{4 }or\texttt{\ "CP"}: conditional power at boundaries
-
-\item \texttt{5} or \texttt{"sf"}: spending function plot (available only if
-\texttt{class(x)=="gsDesign"})
-
-\item \texttt{6} or \texttt{"ASN"} or \texttt{"N"}: expected sample size
-plot
-
-\item \texttt{7"B"} or \texttt{"B-val"} or \texttt{"B-value"}:  B-values at boundaries
-\end{itemize}
-
-\bigskip
-
-
-Further details of the \texttt{print()}, and \texttt{plot()} functions are
-provided in Section~\ref{sec:detailedex}, Detailed Examples, and in the help 
-file for plotting in the gsDesign
-package. Since the plot functions have multiple arguments, the use of these
-arguments is formally specified here.
-
-\bigskip
-
-\texttt{plot.gsDesign(x,plottype=1, main="Default", ylab="Default",
-xlab="Default")}
-
-\texttt{plot.gsProbability(x,plottype=2, xval="Default", main="Default",}
-
-\texttt{\qquad ylab="Default", xlab="Default")}
-
-\bigskip
-
-\begin{itemize}
-\item \texttt{x}: for \texttt{plot.gsDesign()}, this is an object in the
-\texttt{gsDesign} class; for \texttt{plot.gsProbability()}, this is an 
-object in the \texttt{gsProbability} class.
-
-\item \texttt{plottype}: described above.
-
-\item \texttt{theta}: used for \texttt{plottype} equal to 2, 4, or 6; normally 
-defaults are adequate. See the help files for details.
-
-\item \texttt{ses=TRUE}: applies only when \texttt{plottype = 3} and 
-\texttt{class(x) == "gsDesign"};
-indicates that estimated standardized effect size at the boundary
-($\hat{\theta}/\delta$) is to be plotted rather than the actual estimate
-($\hat{\theta}$.
-
-\item \texttt{xval="Default"}: effective only when \texttt{plottype} equals 2 
-or 6.  Appropriately scaled (reparameterized) values for x-axis for power and
-expected sample size graphs; see the help file for details.
-\end{itemize}
-