Generalized additive models have been popular among statisticians and data analysts in multivariate nonparametric regression with non-Gaussian responses including binary and count data. In this paper, a new likelihood approach for fitting generalized additive models is proposed. It aims to maximize a smoothed likelihood. The additive functions are estimated by solving a system of nonlinear inte...

A method of estimating a variety of curves by a sequence of piecewise polynomials is proposed, motivated by a Bayesian model and an appropriate summary of the resulting posterior distribution. A joint distribution is set up over both the number and the position of the knots de®ning the piecewise polynomials. Throughout we use reversible jump Markov chain Monte Carlo methods to compute the poste...

In April 2001 Swiss banks held over CHF 500 billion in mortgages. This important segment accounts for about 63% of all the loan portfolios of Swiss banks. In this paper we restrict our attention to residential mortgages held by private clients, i.e. borrowers who finance their property by the loan and we model the probability distribution of the number of defaults using a non-parametric intensi...

Parsimonious regression models using locational data often yield non-normal, heteroskedastic, and spatially dependent residuals. This manuscript develops a model which simultaneously performs spatial and functional form transformations to mitigate these problems. We apply the model to 11,006 observations on housing prices in Baton Rouge. For these data, the model reduces the interquartile range...

When estimating a generalized additive model, a crucial decision that must be made is the choice of underlying technique that will be used to estimate those transformations that produce the best-fitting model. Data smoothers and universal approximators are two opposing techniques that seem to hold the most promise. ACE (alternating conditional expectations) was developed by Breiman and Friedman...

We estimate a hedonic model of the housing market that includes a spline surface defined on geospatial data (i.e., the longitudes and latitudes of individual dwellings). House price indexes are then obtained by imputing prices for individual dwellings from the hedonic model and then inserting them into the Fisher price index formula. Using data for Sydney, Australia we compare the performance o...

Recursive modeling is an attractive data-analytic tool for studying the relationship between a dependent variable and a collection of predictor variables. In this, it complements methods such as multiple regression, analysis of variance, neural nets, generalized additive models, discriminant analysis and log linear modeling. It is particularly helpful when simple models like regression do not w...

Ethnic variation in mortality and whether this variation can be explained by socioeconomic status are of substantive interest to social epidemiologists. The authors consider the analysis of mortality data for a mixture of majority and minority ethnic groups. Such data are likely to be coarsely cross-classified by age and socioeconomic status and yet, even then, in some cells of this cross-class...

A generalized additive model (GAM) of blue shark, Prionace glauca, catch rates (catch per set) was ®tted to data gathered by National Marine Fisheries Service (NMFS) observers stationed aboard Hawaii-based commercial longline vessels (N ˆ 2010 longline sets) from March 1994 to December 1997. Its coef®cients were then applied to the values of predictor variables, which were also contained in log...

The Tibetan Plateau, known as the “world’s third pole” for its extremely harsh and fragile ecological environment, has attracted great attention because of its sensitivity to global changes. Alpine grassland on the Tibetan Plateau has an important function in the global carbon cycle. Many studies have examined the effects of various environmental factors on biomass distribution. In this study, ...