Modelling and smoothing parameter estimation with multiple quadratic penalties

By Simon N. Wood. 2000

821 citations recorded by the source corpus.

Subfields: Applied Mathematics, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Applied Mathematics

In Statistics and Probability

Citations in the corpus

Spatial modelling using a new class of nonstationary covariance functions (2006, methodology): “One advantage of gam() over Tps() that arises in applications is that gam() allows the inclusion of additional covariates, including additional smooth terms, using an algorithm that can optimize multiple penalty terms (Wood 2000).”
Parsimonious additive models (2007, methodology): “Thus, we find hypothesis tests (Chen, 1993; Chen et al., 1996; Härdle and Korostelev, 1996; Wood, 2000; González–Manteiga and Quintela–del Río, 2002), techniques based on a prediction error estimator (Chambers and Hastie, 1993; Brumback et al., 1999; Shi and Tsai, 1999), as well as Bayesian…”
Splines, knots, and penalties (2010, methodology): “There is much room for improvement here, either using more advanced search methods, like the simplex, or with Newton algorithms, like those Wood (2000) uses for thin-plate splines.”
Winter distribution of blue crab Callinectes sapidus in Chesapeake Bay : application and cross-validation of a two-stage generalized additive model (2005, methodology): “…log abundance, ln(ŷ), at a given location was calculated as the product of Stage I and Stage II (Barry & Welsh 2002): ln(ŷ) = p̂ × ln(µ̂) (3) The flexibility of the response curves was optimized using an iterative method that rewards model fit and penalizes degrees of freedom (Wood 2000).”
Penalized likelihood regression: General formulation and efficient approximation (2002, neutral): “S. N. Wood (2000).”
Decadal-scale changes in the climate and biota of the Pacific sector of the Southern Ocean, 1950s to the 1990s (2005, methodology): “Stronger winds and persistent polynyas would facilitate satisfying these needs, and having fasted considerably longer, quick access to food would be even more critical to the male (see also Kirkwood & Robertson 1997).”
Bayesian P-Splines (2004, methodology): “However, a computationally ef cient algorithm for computing the smoothing parameters was presented by Wood (2000), which seems to work at least for a moderate number of smoothing parameters.”
Recruitment and survival of Northeast Pacific Ocean fish stocks: temporal trends, covariation, and regime shifts (2007, methodology): “For each time series, we tested whether a decadal-scale trend based on smoothing splines provided a significantly better fit than a constant mean model, using an approximate χ2 test (Wood 2000).”
Effects of cold weather on mortality: results from 15 European cities within the PHEWE project. (2008, methodology): “We included in the generalized estimating equations model the basis vector of a cubic regression spline of apparent temperature, obtained by using the mgcv library of R software (24) after defining a city-specific vector of equally spaced knots (1 every 8 C).”
GAMs with integrated model selection using penalized regression splines and applications to environmental modelling (2002, methodology): “The Wood (2000) approach typically converges in 5-15 iterations, with each iteration being O(nq2) in computational cost.”
Bottom-up and top-down controls of walleye pollock ( Theragra chalcogramma) on the Eastern Bering Sea shelf (2006, positive): “A cross-validation approach was used to choose the degree of smoothing for each variable (Wood, 2000), and an approximate F test, in combination with visual examination of the fit, was used to determine whether a relationship was significantly different from linear (Hastie and Tibshirani, 1990).”
Temporal and geographic variability of sardine maturity at length in the northeastern Atlantic and the western Mediterranean (2006, methodology): “)) of fish length class (l ) 12e18 cm and month (mo) (SeptembereAugust, coded with numbers 1e12) within each area and period using Generalized Additive Models with a binomial error distribution and a logit link (Wood, 2000).”