\name{locfit.raw} \alias{locfit.raw} \title{ Local Regression, Likelihood and Density Estimation. } \usage{ locfit.raw(x, y, weights=1, cens=0, base=0, scale=FALSE, alpha=0.7, deg=2, kern="tricube", kt="sph", acri="none", basis=list(NULL), deriv=numeric(0), dc=FALSE, family, link="default", xlim, renorm=FALSE, ev=rbox(), maxk=100, itype="default", mint=20, maxit=20, debug=0, geth=FALSE, sty="none") } \description{ \code{locfit.raw} is an interface to Locfit using numeric vectors (for a model-formula based interface, use \code{\link{locfit}}). Although this function has a large number of arguments, most users are likely to need only a small subset. The first set of arguments (\code{x}, \code{y}, \code{weights}, \code{cens}, and \code{base}) specify the regression variables and associated quantities. Another set (\code{scale}, \code{alpha}, \code{deg}, \code{kern}, \code{kt}, \code{acri} and \code{basis}) control the amount of smoothing: bandwidth, smoothing weights and the local model. Most of these arguments are deprecated - they'll currently still work, but should be provided through the \code{lp()} model term instead. \code{deriv} and \code{dc} relate to derivative (or local slope) estimation. \code{family} and \code{link} specify the likelihood family. \code{xlim} and \code{renorm} may be used in density estimation. \code{ev} specifies the evaluation structure or set of evaluation points. \code{maxk}, \code{itype}, \code{mint}, \code{maxit} and \code{debug} control the Locfit algorithms, and will be rarely used. \code{geth} and \code{sty} are used by other functions calling \code{locfit.raw}, and should not be used directly. } \arguments{ \item{x}{ Vector (or matrix) of the independent variable(s). Can be constructed using the \code{lp()} function. } \item{y}{ Response variable for regression models. For density families, \code{y} can be omitted. } \item{weights}{ Prior weights for observations (reciprocal of variance, or sample size). } \item{cens}{ Censoring indicators for hazard rate or censored regression. The coding is \code{1} (or \code{TRUE}) for a censored observation, and \code{0} (or \code{FALSE}) for uncensored observations. } \item{base}{ Baseline parameter estimate. If provided, the local regression model is fitted as \eqn{Y_i = b_i + m(x_i) + \epsilon_i}, with Locfit estimating the \eqn{m(x)} term. For regression models, this effectively subtracts \eqn{b_i} from \eqn{Y_i}. The advantage of the \code{base} formulation is that it extends to likelihood regression models. } \item{scale}{ Deprecated - see \code{\link{lp}()}. } \item{alpha}{Deprecated - see \code{\link{lp}()}. A single number (e.g. \code{alpha=0.7}) is interpreted as a nearest neighbor fraction. With two componentes (e.g. \code{alpha=c(0.7,1.2)}), the first component is a nearest neighbor fraction, and the second component is a fixed component. A third component is the penalty term in locally adaptive smoothing. } \item{deg}{ Degree of local polynomial. Deprecated - see \code{\link{lp}()}. } \item{kern}{ Weight function, default = \code{"tcub"}. Other choices are \code{"rect"}, \code{"trwt"}, \code{"tria"}, \code{"epan"}, \code{"bisq"} and \code{"gauss"}. Choices may be restricted when derivatives are required; e.g. for confidence bands and some bandwidth selectors. } \item{kt}{ Kernel type, \code{"sph"} (default); \code{"prod"}. In multivariate problems, \code{"prod"} uses a simplified product model which speeds up computations. } \item{acri}{Deprecated - see \code{\link{lp}().}} \item{basis}{User-specified basis functions.} %See \code{\link{lfbas}} for more details on this argument.} \item{deriv}{ Derivative estimation. If \code{deriv=1}, the returned fit will be estimating the derivative (or more correctly, an estimate of the local slope). If \code{deriv=c(1,1)} the second order derivative is estimated. \code{deriv=2} is for the partial derivative, with respect to the second variable, in multivariate settings. } \item{dc}{ Derivative adjustment. } \item{family}{ Local likelihood family; \code{"gaussian"}; \code{"binomial"}; \code{"poisson"}; \code{"gamma"} and \code{"geom"}. Density and rate estimation families are \code{"dens"}, \code{"rate"} and \code{"hazard"} (hazard rate). If the family is preceded by a \code{'q'} (for example, \code{family="qbinomial"}), quasi-likelihood variance estimates are used. Otherwise, the residual variance (\code{\link{rv}}) is fixed at 1. The default family is \code{"qgauss"} if a response \code{y} is provided; \code{"density"} if no response is provided. } \item{link}{ Link function for local likelihood fitting. Depending on the family, choices may be \code{"ident"}, \code{"log"}, \code{"logit"}, \code{"inverse"}, \code{"sqrt"} and \code{"arcsin"}. } \item{xlim}{ For density estimation, Locfit allows the density to be supported on a bounded interval (or rectangle, in more than one dimension). The format should be \code{c(ll,ul)} where \code{ll} is a vector of the lower bounds and \code{ur} the upper bounds. Bounds such as \eqn{[0,\infty)} are not supported, but can be effectively implemented by specifying a very large upper bound. } \item{renorm}{Local likelihood density estimates may not integrate exactly to 1. If \code{renorm=T}, the integral will be estimated numerically and the estimate rescaled. Presently this is implemented only in one dimension. } \item{ev}{ The evaluation structure, \code{\link{rbox}()} for tree structures; \code{\link{lfgrid}()} for grids; \code{\link{dat}()} for data points; \code{\link{none}()} for none. A vector or matrix of evaluation points can also be provided, although in this case you may prefer to use the \code{\link{smooth.lf}()} interface to Locfit. Note that arguments \code{flim}, \code{mg} and \code{cut} are now given as arguments to the evaluation structure function, rather than to \code{locfit.raw()} directly (change effective 12/2001). } \item{maxk}{ Controls space assignment for evaluation structures. For the adaptive evaluation structures, it is impossible to be sure in advance how many vertices will be generated. If you get warnings about `Insufficient vertex space', Locfit's default assigment can be increased by increasing \code{maxk}. The default is \code{maxk=100}. } \item{itype}{ Integration type for density estimation. Available methods include \code{"prod"}, \code{"mult"} and \code{"mlin"}; and \code{"haz"} for hazard rate estimation problems. The available integration methods depend on model specification (e.g. dimension, degree of fit). By default, the best available method is used. } \item{mint}{ Points for numerical integration rules. Default 20. } \item{maxit}{ Maximum iterations for local likelihood estimation. Default 20. } \item{debug}{If > 0; prints out some debugging information.} \item{geth}{Don't use!} \item{sty}{ Deprecated - see \code{\link{lp}()}. } } \value{ An object with class "locfit". A standard set of methods for printing, ploting, etc. these objects is provided. } \keyword{smooth}