B. D. Ripley's research while affiliated with University of Oxford and other places

Publications (92)

Article
We introduce a method of estimating disease prevalence from case-control family study data. Case-control family studies are performed to investigate the familial aggregation of disease; families are sampled via either a case or a control proband, and the resulting data contain information on disease status and covariates for the probands and their...
Book
Ripley brings together two crucial ideas in pattern recognition: statistical methods and machine learning via neural networks. He brings unifying principles to the fore, and reviews the state of the subject. Ripley also includes many examples to illustrate real problems in pattern recognition and how to overcome them.
Article
Likert attitude data consist of responses to favorable and unfavorable statements about an entity, where responses fall into ordered categories ranging from disagreement to agreement. Social science and marketing researchers frequently use data of this type to measure attitudes toward an entity such as a policy or product. We focus on data on Ameri...
Article
Likert attitude data consist of responses to favorable and unfavorable statements about an entity, where responses fall into ordered categories ranging from disagreement to agreement. Social science and marketing researchers frequently use data of this type to measure attitudes toward an entity such as a policy or product. We focus on data on Ameri...
Chapter
One sense of “computer-intensive” statistics is just statistical methodology that makes use of a large amount of computer time. (Examples include the bootstrap, jackknife, smoothing, image analysis, and many uses of the EM algorithm.) However, the term is usually used for methods that go beyond the minimum of calculations needed for an illuminating...
Article
Traditionally a 'model' is a family of probability distributions for the observed data parametrized by a set of parameters (of fixed and finite dimension), but it is often helpful to consider all the models considered as subsets of one model, as well as some even larger models used in 'over-fitting' as part of the validation process. Traditional di...
Chapter
Classification is an increasingly important application of modern methods in statistics. In the statistical literature the word is used in two distinct senses. The entry (Hartigan, 1982) in the original Encyclopedia of Statistical Sciences uses the sense of cluster analysis discussed in Section 11.2. Modern usage is leaning to the other meaning (Ri...
Chapter
In linear regression the mean surface is a plane in sample space; in non-linear regression it may be an arbitrary curved surface but in all other respects the models are the same. Fortunately the mean surface in most non-linear regression models met in practice will be approximately planar in the region of highest likelihood, allowing some good app...
Chapter
Multivariate analysis is concerned with datasets that have more than one response variable for each observational or experimental unit. The datasets can be summarized by data matrices X with n rows and p columns, the rows representing the observations or cases, and the columns the variables. The matrix can be viewed either way, depending on whether...
Article
In functional magnetic resonance imaging statistical analysis there are problems with accounting for temporal autocorrelations when assessing change within voxels. Techniques to date have utilized temporal filtering strategies to either shape these autocorrelations or remove them. Shaping, or "coloring," attempts to negate the effects of not accura...
Chapter
We collect together several ways to handle linear and non-linear models with random effects, possibly as well as fixed effects.
Chapter
Statistics is fundamentally about understanding data. We start by looking at how data are represented in S, then move on to importing, exporting and manipulating data.
Chapter
Statisticians1 often under-estimate the usefulness of general optimization methods in maximizing likelihoods and in other model-fitting problems. Not only are the general-purpose methods available in the S environments quick to use, they also often outperform the specialized methods that are available. A lot of the software we have illustrated in e...
Book
A guide to using S environments to perform statistical analyses providing both an introduction to the use of S and a course in modern statistical methods. The emphasis is on presenting practical problems and full analyses of real data sets.
Article
There are many ways to detect activation patterns in a time series of observations at a single voxel in a functional magnetic resonance imaging study. The critical problem is to estimate the statistical significance, which depends on the estimation of both the magnitude of the response to the stimulus and the serial dependence of the time series an...
Chapter
The new S engine introduced a very different approach to classes, although backwards compatibility is provided for the classes as described in Chapter 4. The current S-PLUS systems rely very heavily on this backwards compatibility; at present new-style classes are used only at a low level and in the time-series manipulation software. The definitive...
Chapter
This chapter provides a reprise of the material introduced in Chapters 2 and 3 of MASS. Chapter 3 introduces more advanced language concepts that are important for programming: there are also class-oriented features which we discuss in Chapters 4 and 5.
Chapter
In this chapter we consider the details which are not normally important in interactive use of S, and some more formal aspects of the language.
Chapter
Most S-PLUS programmers find early in their careers that they have written code which exhausts the physical memory (RAM) available to the S-PLUS process, and then proceeds to spend almost all its time in allocating virtual memory. Such code can reduce the fastest workstation to page thrashing, which will reduce severely the size of problem which ca...
Chapter
A very important and powerful feature of the S environment is that it is not restricted to functions written in the S language but may load and use compiled routines written in C or Fortran, or perhaps other languages.1 Moreover there are ways in which such externally written and compiled routines may communicate with the S session and make use of...
Chapter
S-PLUS 4.x for Windows introduced a new way to program menus and dialogs. In essence the system programs the Axum engine on which the 4.x GUI is based. Even those who much prefer the command-line interface (including the authors) may find they need to develop a graphical user interface to their functions to enable other people to make use of them.
Chapter
This chapter covers the class-oriented features of the old S engine and of R. Most of these exist for backward compatibility in the new S engine (see Section 5.4), but the features introduced in the next chapter are preferred for new projects in that system.
Chapter
In this chapter we consider the tools available for the process of software development in the S language. Such tools are often a matter of personal taste so many are available, and our aim is to cover all the possibilities at fairly shallow level.
Chapter
In this chapter we cover a number of topics from classical univariate statistics plus some modern versions.
Chapter
Statistics is fundamentally concerned with the understanding of structure in data. One of the effects of the information-technology era has been to make it much easier to collect extensive datasets with minimal human intervention. Fortunately the same technological advances allow the users of statistics access to much more powerful ‘calculators’ to...
Chapter
S-PLUS has a ‘Modern Regression Module’ which contains functions for a number of regression methods. These are not necessarily non-linear in the sense of Chapter 8, which refers to a non-linear parametrization, but they do allow nonlinear functions of the independent variables to be chosen by the procedures. The methods are all fairly computer-inte...
Chapter
Multivariate analysis is concerned with datasets that have more than one response variable for each observational or experimental unit. The datasets can be summarized by data matrices X with n rows and p columns, the rows representing the observations or cases, and the columns the variables. The matrix can be viewed either way, depending on whether...
Chapter
There are now a large number of books on time series. Our philosophy and notation are close to those of the applied book by Diggle (1990) (from which some of our examples are taken). Brockwell & Davis (1991) and Priestley (1981) provide more theoretical treatments, and Bloomfield (1976) and Priestley are particularly thorough on spectral analysis....
Chapter
S-PLUS provides comprehensive graphics facilities, from simple facilities for producing common diagnostic plots by plot (object) to fine control over publication-quality graphs. In consequence, the number of graphics parameters is huge. In this chapter, we build up the complexity gradually. Most readers will not need the material in Section 3.4, an...
Article
If we view neural nets as a class of statistical models with highdimensional parameters, we can consider how to apply the ideas of statistical theory, in particular ideas for model choice and the concepts of predictive Bayesian inference. It turns out that these ideas give considerable insight, and enable us to find more powerful solutions with red...
Book
S-Plus is a powerful environment for statistical and graphical analysis of data. It provides the tools to implement many statistical ideas which have been made possible by the widespread availability of workstations having good graphics and computational capabilities. This book is a guide to using S-Plus to perform statistical analyses and provides...
Chapter
Statistics is fundamentally concerned with the understanding of structures in data. One of the effects of the information-technology era has been to make it much easier to collect extensive datasets with minimal human intervention. Fortunately the same technological advances allow the users of statistics access to much more powerful ‘calculators’ t...
Chapter
In linear regression the mean surface in sample space is a plane; in non-linear regression it may be an arbitrary curved surface but in all other respects the models are same. Fortunately in practice the mean surface in most non-linear regression models will be approximately planar in the region of highest likelihood, allowing some good approximati...
Article
ashing powders (Carstensen, 1992). In all these of tasks there is a predefined set of classes of patterns which might be presented, and the task is to classify a future pattern as one of these classes. Such tasks are called classification or supervised pattern recognition 1 . Clearly someone had to determine the classes in the first place, and seek...
Article
Choosing the architecture of a neural network is one of the most important problems in making neural networks practically useful, but accounts of applications usually sweep these details under the carpet. How many hidden units are needed? Should weight decay be used, and if so how much? What type of output units should be chosen? And so on. We addr...
Article
Feed‐forward neural networks are now widely used in classification problems, whereas non‐linear methods of discrimination developed in the statistical field are much less widely known. A general framework for classification is set up within which methods from statistics, neural networks, pattern recognition and machine learning can be compared. Neu...
Chapter
S-PLUS provides comprehensive graphics facilities, from simple facilities for producing common diagnostic plots by plot (object) to fine control over publication-quality graphs. In consequence, the number of graphics parameters is huge. In this chapter, we build up the complexity gradually. Most readers will not need to go beyond the first 3 sectio...
Chapter
S is a language for the manipulation of objects. It aims to be both an interactive language (like, for example, a Unix shell language) as well as a complete programming language with some convenient object-oriented features. In this chapter we shall be concerned with the interactive language, and hence certain language constructs used mainly in pro...
Article
The use of tree-based models will be relatively unfamiliar to statisticians, although researchers in other fields have found trees to be an attractive way to express knowledge and aid decision-making. Keys such as Figure 10.1 are common in botany and in medical decision-making, and provide a way to encapsulate and structure the knowledge of experts...
Chapter
Survival analysis is not part of S, but has been added to S-PLUS based on functions written by Terry Therneau (Mayo Foundation) and available as survival2 code from statlib (see Appendix D for further information.) The functions in survival3 were released in mid-1992. They are not part of S-PLUS 3.2, but are scheduled to be included in late 1994. A...
Article
Neural networks are one of a class of classifiers which construct a nonlinear function from inputs to targets. There are a series of questions common to all members of the class, including how best to use the outputs to classify, how to fit the class of functions and also how to choose between classes. These questions are explored via some theory a...
Chapter
Multivariate analysis is concerned with datasets which have more than one response variable for each observational or experimental unit. The datasets can be summarized by data matrices X with n rows and p columns, the rows representing the observations or cases, and the columns the variables. The matrix can be viewed either way, depending whether t...
Chapter
Outliers are sample values which cause surprise in relation to the majority of the sample. This is not a pejorative term; outliers may be correct, but they should always be checked for transcription errors. They can play havoc with standard statistical methods, and many robust and resistant methods have been developed since 1960 to be less sensitiv...
Chapter
The S language is both an interactive language and a language for adding new functions to the S system. It is a complete programming language with control structures, recursion and a useful variety of data types. The S environment provides many functions to handle standard operations, but most users need occasionally to write new functions. This ch...
Chapter
S-PLUS has a ‘Modern Regression Module’ which contains functions for a number of regression methods. These are not necessarily non-linear in the sense of Chapter 9, which refers to a non-linear parametrization, but they do allow nonlinear functions of the independent variables to be chosen by the procedures. The methods are all fairly computer-inte...
Article
Feed-forward neural networks are now widely used in classification problems, whereas nonlinear methods of discrimination developed in the statistical field are much less widely known. A general framework for classification is set up within which methods from statistics, neural networks, pattern recognition and machine learning can be compared. Neur...
Chapter
Linear models form the core of classical statistics, and S provides extensive facilities to fit and manipulate them. These work with a version of the Wilkinson-Rogers syntax (Wilkinson & Rogers, 1973) for specifying models which we discuss in the Section 6.2, and which is also used for generalized linear models, models for survival analysis and tre...
Article
Feed-forward neural networks—also known as multi-layer perceptrons—are now widely used for regression and classification. In parallel but slightly earlier, a family of methods for flexible regression and discrimination were developed in multivariate statistics, and tree-induction methods have been developed in both machine learning and statistics....
Chapter
There are now a large number of books on time series. Our philosophy and notation are close to those of the applied book by Diggle (1990) (from which some of our examples are taken). Brockwell and Davis (1991) and Priestley (1981) provide more theoretical treatments, and Bloomfield (1976) and Priestley are particularly thorough on spectral analysis...
Chapter
Generalized linear models (GLMs) extend linear models to accommodate both non-normal response distributions and transformations to linearity. (We will assume that Chapter 6 has been read before this chapter.) The essay by Firth (1991) gives a good introduction to GLMs; the comprehensive reference is McCullagh & Neider (1989).
Chapter
In this chapter we cover a number of topics from classical univariate statistics. Many of the functions used are S-PLUS extensions to S.
Chapter
In linear regression the mean surface in sample space is a plane. In non-linear regression the mean surface may be an arbitrary curved surface but in other respects the models are similar. In practice the mean surface in most non-linear regression models will be approximately planar in the region(s) of high likelihood allowing good approximations b...
Chapter
Statistics is fundamentally concerned with the understanding of structures in data. One of the effects of the information-technology era has been to make it much easier to collect extensive datasets with minimal human intervention. Fortunately the same technological advances allow the users of statistics access to much more powerful ‘calculators’ t...
Article
Optical astronomers now normally collect digital images by means of charge-coupled device detectors, which are blurred by atmospheric motion and distorted by physical noise in the detection process. We examine Bayesian procedures to clean such images using explicit models from spatial statistics for the underlying structure, and compare these metho...
Article
In this paper we present a Bayesian method to deconvolve images when the location of the objects in the image is known in advance. This knowledge of location is incorporated into the prior model via a labeling process. An iterative method is proposed to find the maximum a posteriori estimator of the image. The method was tested on both synthetic im...
Article
Full-text available
Bayesian methods and spatial stochastic processes are used here in the deconvolution of images of galaxies. Under very simple but realistic prior assumptions about the true underlying image of a galaxy the Bayesian framework is put to work. The method is tested in CCD images of extragalactic objects of different morphological types and an analysis...
Conference Paper
Considers Bayesian methods and spatial stochastic processes applied to the deconvolution of images of planets. Under simple but realistic prior assumptions about the true underlying image of a planet the Bayesian framework is put to work. The method has been tested on CCD images of Jupiter
Article
Problems of classification form a large part of the literature on pattern recognition, and of this a large subset is devoted to problems of classification and recognition in image data. After a brief introduction to the area, we describe two problems which we are currently working on, both involving the application of formal statistical methods to...
Article
Formal Bayesian methods have only a little history in astronomical applications, yet they have recently become the favorite methodology for statisticians studying image analysis. The “prior distributions” used are spatial stochastic processes which aim to encapsulate the relevant features of the images which are known from past experience. We descr...
Chapter
Monte Carlo methods have a long history in spatial statistics, and have often been used very effectively to sidestep problems of analytical or computational intractability. On the other hand, bootstrap and other non-parametric methods have made no impact and are rarely considered. The reasons are immediate but often overlooked. The author was once...
Chapter
We describe in this work how the Bayesian paradigm can be applied to a deconvolution problem in optical astronomy. The use of robust statistics in this process is also discussed.
Article
Much recent work in statistical image analysis has been concerned with `cleaning' images by a bayesian statistical analysis incorporating a prior model, which reflects the spatial structure of the image. In almost all cases this has involved a description of the image at pixel level. In this paper we take the process further, and develop a spatial...
Article
Much of the informal discussion at the Workshop concerned the merits of different pseudorandom number generators. Here we record some comments based on comparing generators across a wide range of machines.
Article
The standard methods of generating sample from univariate distributions often become hopelessly inefficient when applied to realizations of stochastic processes. Iterative methods are not widely known amongst statisticians, but some are standard practice in statistical physics and chemistry. The methods are surveyed and compared, with particular re...
Article
It is well known that the theoretical intensity at a point (x,y) from an astronomical image is given $$ \tilde Z(x,y) = \iint {S(x - x',y - y')h(x',y')dx'dy'} $$ (1) where S(,) represents the true underlying intensity and h(,) is the point spread function (psf)-However, when we consider discrete samples on a rectangular grid, the intensity measured...
Article
Simulation is a widely used methodology for queueing systems. Its superficial simplicity hides a number of pitfalls which are not all as well known as they should be. In particular simulation experiments need careful design and analysis as well as good presentations of the results. Even the elements of simulation such as the generation of arrival a...
Article
More and more problems are being tackled by simulation as large computing costs per hour approach those of mathematicians' time. Abuses of simulation arise from ignorance or careless use of little understood procedures, and some of the fundamental tools of the subject are much less well understood than commonly supposed. This is illustrated here by...
Chapter
Point processes are models for point-like objects such as trees in a forest and earthquake occurrences in space-time. They are used to produce precise descriptions of point patterns. The basic types of pattern are described together with informative graphical summaries.
Article
Regression techniques are commonly applied to compare two analytical methods at several concentrations and to test the biases of one method relative to another. However, regression is strictly applicable only when one method is without error, for example in comparisons with reference materials. A regression-like technique, maximum-likelihood fittin...
Chapter
Statistics has been applied to ecological problems involving spatial patterns for most of this century. Even in the 1950’s quite specialised methods had been developed for detecting “scale” in grassland and to census mobile animal populations (especially game). After a general discussion this paper concentrates on point patterns and their analysis...
Article
The Tools Models Simulation as Experimentation Examples Literature Convention
Article
This first correspondent's paper on spatial statistics surveys recent developments in the field, updating the author's monograph Spatial Statistics. The main themes are edge effects, multitype point processes and computation. /// Dans cette première communication de statistique spatiale on présente les développements récents du sujet. Elle met à jo...
Article
Spatial statistics is a recent and graphical subject which is ideally suited to implementation in S; S itself includes one spatial interpolation method, akima, and loess which can be used for two-dimensional smoothing, but the specialist methods of spatial statistics have been added and are given in our library spatial. The main references for spat...
Book
Basic ParametersNearest-Neighbor Methods Second MomentsModelsComparative StudiesExamples
Article
(1) Spectral analysis is a relatively untried method for the analysis of data from a line of contiguous quadrats. Conventional block-size analyses are shown to be related to square waves. In spectral analysis square waves are replaced by sine waves. (2) These methods and Mead’s test are compared with conventional methods, using artificial and field...
Article
Data Mining is a modern buzzword for finding (useful) information from large (often massive) databases. If has also been an opportunity for the re-invention (or re-marketing) of many ideas from statistics and machine learning, and the marketing of many commercial programs for 'data mining'. In the words of Witten & Franke (2000, p. 26) What's the d...

Citations

... Traditional algorithms have the disadvantage of requiring a great deal of timeconsuming feature engineering and manual feature extraction. Artificial neural networks are a class of flexible nonlinear regression models ( Kar, 2016 ;Mastinu et al., 2007 ;Ripley, 1994 ;Sarle, 1994 ). A key characteristic of ANN consists of imitating the biological functioning of the human brain and self-organizing its internal processes without external interference. ...
... Because it features "shortcut connections," RN differs from the canonical MLP architecture in that it mitigates the problem of degradation in the event of numerous layers (He et al., 2015). Although the use of shortcut connections is not new in the literature (Venables & Ripley, 1999), He et al. (2015) proposed that identity mapping be used instead of any other nonlinear transformation. The simplest building unit of the ResNet architecture is depicted in Fig. 1. ...
... Há agora alguns livros que descrevem como usar o R para análise de dados e estatísticas, bem como documentação para o S e o S-Plus, que podem ser usados juntamente com o R, mantendo as diferenças entre eles(VENABLES, 2004).Muitas pessoas usam o R como um sistema estatístico e, sendo assim, muitas estatísticas modernas foram implementadas. Algumas delas são construídas na base R de desenvolvimento e outras são fornecidas como "pacotes". ...
... The two-step approach described above ( Figure 2) is general; additional classification methods can be included, and the approach can be applied to any (natural landscape) classification task. Model selection and variable explanation were implemented in R [51]. ...
... All analyses were performed in R [47] using the "chisq.test" and "glm" functions from the STATS package for Logit Models [48]. The results from chi-square and odds ratio statistical analyses are provided in Supplementary Table 1 and Table 2, respectively. ...
... The following R packages were necessary to conduct our analyses: vegan [33], betapart [34], MASS [35], Peptides [36], datatable [37], and ggplot2 [38]. Our catchment map (Fig. 1) was constructed using data available from the Environment Agency (UK) Catchment Data Explorer portal [39] and manipulated using the sf package in R [40]. ...
... Habitat suitability maps were calculated using 12 different algorithms. These algorithms are: Bioclim [33,48,49], Domain [76], Generalized Linear Models (GLM) [77,78], Generalized Additive Models (GAM) [78,79], Multivariate Adaptive Regression Splines (MARS) [80][81][82], Flexible Discriminant Analysis (FDA) [83][84][85], Classification Tree Analysis (CTA) [86], Artificial Neural Network (ANN) [87,88], Random Forest (RF) [89,90], Support Vector Machine (SVM) [91][92][93], Maximum Entropy (Maxent) [24,35,94], and Kernel Density Estimation (KDE) [95,96]. All algorithms were used in R [97]. ...
... We tested reliability of lens isotopic dietary values to classify individuals into their correct rearing habitat using a linear discriminant analysis (LDA) utilizing jackknifed cross validation in the mass package in R version 4.0.2 (Venables et al., 2002). Lastly, individual laminae from a single juvenile salmon from the floodplain were analysed for δ 13 C, δ 15 N and δ 34 S to test the feasibility of reconstructing diets over smaller temporal and spatial scales of habitat use. ...
... " Halliday (1967) adds that given information is often represented anaphorically, by means of reference (pronominals and demonstratives), substitutes (words like one and do), and ellipsis (no realisation in the text). Moreover, in English sentences, usually the portion bearing given information precedes the portion conveying new information (Quirk, Greenbaum, Leech, & Svartvik, 1972; Prince, 1978; Chafe, 1979; Kuno, 1980; Fries, 1983). The portion that bears the given information is often the complete subject, and the portion that bears the new information is often the complete predicate. ...