Gaussian additive model Gram-Schmidt method Hilbert space Local-likelihood estimation Los Angeles air-pollution data Newton-Raphson algorithm Newton-Raphson iterations Newton-Raphson update ace criterion additive predictor additive proportional hazards additive spline model adjusted dependent variable regression asymptotic variance-stabilizing transformation backfitting algorithm backfitting solutions balanced additive bandwidth selection basis function bin smoother binary data bootstrap approximation boundary knots canonical link cardinal-splines conditional expectation conditional expectation operators constant term continuous response variable cubic smoothing splines cubic smoothing-spline operator matrix cubic splines cubic-spline basis functions delta algorithms dummy variable equivalent kernel equivalent kernels estimating equations evaluated-splines exact concurvity expected log-likelihood exponential family models fitted function fitted values formula language generalized additive models generalized cross-validated deviance generalized residual hazard function hierarchical model isotonic regression jackknifed fit linear effect linear predictor linear scatterplot smoother linear system local-scoring algorithm log-linear models logistic additive model logistic regression main effects missing data modelling interactions moving average multi-predictor smoothers multinomial likelihood multiple regression natural cubic spline natural parameter nearest-neighbour smoother non-DOS computers nonhierarchical model null hypothesis observation weights orthogonal polynomials ozone concentration data parametric fitting partial residuals partial-residual plots penalized least squares penalized log-likelihood piecewise cubic polynomials piecewise polynomials pointwise standard-error bands pointwise standard-error curves posterior covariance predictive ability prior covariance projection-type smoothers pseudo additive models reproducing-kernel Hilbert-spaces resistant additive ridge regression scale estimate scale parameter score equations seasonal operators shrinking smoothers slicing regression smoothing parameter smoothing parameter selection smoothing spline span selection specialized local-scoring algorithm standard-error bands step size optimization stepwise-additive methods surface smoother symmetric smoother matrices tensor product bases tidwell method time series trend component unweighted running-lines smoother weight function weighted additive-fit operator weighted cubic smoothing spline weighted least-squares fit weighted smoothers Cox model Fourier coefficients S functions adaptive techniques additive predictor adjusted dependent variable alternating algorithm binary data comparing models cross-validated deviance diagonal elements fitted value global confidence band influential points linear filter local-scoring algorithm locally-weighted running-lines logistic regression matched sets maximum likelihood monotone transformations natural splines nearest neighbours optimal rate ozone data partial residual penalized least-squares criterion pointwise standard-error bands posterior covariance posterior mean proportional-odds model resistant algorithm running median scatterplot smoother scatterplot smoothing seasonal effect semi-parametric model smoother matrix smoothing parameter selection smoothing-spline matrix standard-error bands survival data symmetric nearest neighbourhood target value trend smoother weight matrix weighted additive model asymptotic bias bootstrap sample cubic smoothing spline equivalent kernel estimating equations expected log-likelihood information matrix kernel smoother least-squares line local-likelihood estimation matched case-control data modified backfitting algorithm multiple linear regression nonlinear smoothers partial likelihood posterior distribution running-lines smoother seasonal smoother stl procedure surface smoothers time series Kullback-Leibler distance additive model backfitting algorithm concurvity space equivalent kernel local-scoring algorithm locally-weighted running-lines seasonal component time series transfer function Bayesian model conditional likelihood cubic smoothing spline generalized additive model model selection orthogonal projection smoother matrix smoothing parameter unique solution backfitting converges interior knots kernel smoothers maximal correlation canonical correlation running mean smoothing splines adjusted dependent variable exponential family generalized linear model proportional-hazards model response transformation optimal transformations link function starting functions backfitting algorithm ace algorithm smoothing parameter backfitting algorithm estimating equations {\em ACE and Correspondence analysis} {\em ACE and canonical correlation} {\em Atmospheric ozone concentration} {\em Automatic backfitting} {\em Average Derivative Estimation} {\em CART} software {\em Computation of the \GCV\ statistic} {\em Delta method} {\em Diabetes data} {\em Efficient kernel smoothing ^{Silverman (1982)}, {\em Generalized cross-validation (GCV)} {\em Hanning} {\em Kriging} {\em Kyphosis data} {\em M-estimate} approaches {\em M-estimation for regression} {\em M-estimation} {\em Mallow's $C_p$} {\em Semi-parametric regression ^^{Green, P.J.}^^{Jennison, {\em Slicing regression} {\em TURBO} paper {\em Twicing} {\em Universal Kriging} {\em Updating formula for running-line smooth} {\em Warm cardioplegia data} {\em additive model} {\em additive predictor} {\em additive} predictor {\em adjusted dependent variable regression} {\em asymptotic variance stabilizing transformation} {\em backfitting algorithm} {\em backfitting} {\em calendar} effects {\em canonical correlation} {\em canonical link} {\em centered} smoother {\em collinearity} {\em complimentary log-log} {\em concurvity space} {\em concurvity} {\em convolution} {\em curse of dimensionality} {\em degrees of freedom} {\em delta algorithm} {\em deviance} {\em digital filter}) {\em effective number of parameters} {\em effective} dimension {\em equivalent degrees of freedom} {\em equivalent kernel} {\em estimating} equations {\em expected} log-likelihood {\em frequency response functions} {\em fundamental tradeoff between bias and variance} {\em generalized additive models} {\em generalized additive model} {\em generalized linear models} {\em hat} matrix {\em hierarchical} {\em impulse response function} {\em interaction} {\em leverage points} {\em linear predictor} {\em link function} {\em local scoring} {\em local-likelihood} estimation {\em local-scoring procedure} {\em loess} {\em loess}) {\em logit} {\em low pass} {\em matched sets} {\em maximal correlation} {\em missing at random} {\em nonparametric} nature {\em odds-ratio} {\em optimal transformations for correlation} {\em optimal transformations for regression} {\em optimal transformations} {\em partial likelihood} {\em powering up} {\em probit} {\em profile log-likelihood} {\em pseudo additive models} {\em pseudo smoothers} {\em regression smoothers} {\em representers of evaluation} {\em resubstitution prediction error} {\em ridge regression} {\em scatterplot smoother} {\em seasonal} smoother {\em semi-parametric} model {\em shrinking} {\em shrinking} smoothers {\em simple Kriging} {\em smoother matrix} {\em smoothing parameter} {\em splitting} {\em state-space} approach {\em supersmoother} {\em tensor product} {\em thin-plate spline} {\em ties} {\em trading day} {\em transfer function} {\em transformation} {\em trend} smoother {\em tri-cube} weight {\em twicing} {\em{BOX-TIDWELL}}} {\em{BRUTO}}} {\em{STEP-ADDITIVE}}} {\em{TURBO}}}