Search

We enjoyed reading Professor Efron's (Brad) paper just as much as we enjoyed listening to his June 2019 lecture in Leiden. Interestingly, Efron puts the pure predictive algorithms to the test in a scenario where the sample size is extremely moderate by today's standards ( I n i = 102). In the knockoff framework, one can use I any i black-box (any predictive model) to identify any kind possible association - linear or not - between a large set of explanatory variables HT ht and an outcome I Y i . This framework rigorously tests whether the conditional distribution HT ht actually depends on any of the variables I X SB j sb i without assuming any model for how I X i and I Y i are connected (Candès et al. 2018). [Extracted from the article]

Summary For a Bayesian agent with beliefs about the relationship between covariates and potential outcomes, deterministically selecting an assignment that yields optimal covariate balance rationally dominates randomisation. However, randomisation—by enabling control over the probabilities of erroneous causal conclusions due to unknown covariate imbalances—offers insurance against the possibility that an agent's beliefs may be misleading. For the most part, such rational justifications for optimum assignment have presupposed the framework of Bayesian inference, while such epistemic justifications for randomisation have presupposed the framework of significance testing. In this paper, I build on a conception of balance that seems inextricable from the significance testing framework, Fisherian balance, to show that it implies an analogous epistemic justification for randomisation within the framework of Bayesian inference. Consequently, for the choice between optimum and random assignment, this paper shows that epistemic justifications need not be wedded to significance testing nor must Bayesian inference be wedded to rational justifications. [ABSTRACT FROM AUTHOR]

Summary Unemployment rate estimates for small areas are used to efficiently support the distribution of services and the allocation of resources, grants and funding. A Fay–Herriot type model is the most used tool to obtain these estimates. Under this approach out‐of‐sample areas require some synthetic estimates. As the geographical context is extremely important for analysing local economies, in this paper, we allow for area random effects to be spatially correlated. The spatial model parameters are estimated by a marginal likelihood method and are used to predict in‐sample as well as out‐of‐sample areas. Extensive simulation experiments are used to assess the impact of the auto‐regression parameter and of the rate of out‐of‐sample areas on the performance of this approach. The paper concludes with an illustrative application on real data from the Italian Labour Force Survey in which the estimation of the unemployment rate in each Local Labour Market Area is addressed. [ABSTRACT FROM AUTHOR]

Summary Probability weights have been widely used in addressing selection bias arising from a variety of contexts. Common examples of probability weights include sampling weights, missing data weights, and propensity score weights. Frequency weights, which are used to control for varying variabilities of aggregated outcomes, are both conceptually and analytically different from probability weights. Popular software such as R, SAS and STATA support both types of weights. Many users, including professional statisticians, become bewildered when they see identical estimates, but different standard errors and p$$ p $$‐values when probability weights are treated as frequency weights. Some even completely ignore the difference between the two types of weights and treat them as the same. Although a large body of literature exists on each type of weights, we have found little, if any, discussion that provides head‐to‐head comparisons of the two types of weights and associated inference methods. In this paper, we unveil the conceptual and analytic differences between the two types of weights within the context of parametric and semi‐parametric generalised linear models (GLM) and discuss valid inference for each type of weights. To the best of our knowledge, this is the first paper that looks into such differences by identifying the conditions under which the two types of weights can be treated the same analytically and providing clear guidance on the appropriate statistical models and inference procedures for each type of weights. We illustrate these considerations using real study data. [ABSTRACT FROM AUTHOR]

Summary: This paper briefly reviews the recent research in matrix‐variate time series analysis, discusses some new developments, especially for seasonal time series, and demonstrates some applications. A general matrix autoregressive moving‐average model is introduced. The paper narrates a simple approach for understanding the model, identifiability issues, and estimation. Real examples are used to illustrate the theory. [ABSTRACT FROM AUTHOR]

A all for papers on energy statistics, to be published in the special issue of the journal is presented.

Summary In this paper, we begin by reviewing the optional randomised response technique estimator (ORRTE) developed by Chaudhuri and Mukerjee for estimating the proportion of a sensitive characteristic in a population. We show that their estimator is unbiased and has smaller variance than the Warner's estimator. Then we make an attempt at developing an optimised optional randomised response technique estimator (OORRTE). The proposed OORRTE is shown to be more efficient than the ORRTE. Findings from simulation studies are discussed and interpreted for various situations. Sample sizes for the Warner's estimator, the ORRTE and the OORRTE are computed based on power analysis introduced by Ulrich, Schroter, Striegel and Simon. Finally, we include an application to real data on COVID‐19 by considering it to be partially sensitive variable; that is, sensitive to some but not to others. The data used are included in the paper and the R‐codes used in the simulation study are documented in online material. [ABSTRACT FROM AUTHOR]

Summary: Crowdsourcing methods facilitate the production of scientific information by non‐experts. This form of citizen science (CS) is becoming a key source of complementary data in many fields to inform data‐driven decisions and study challenging problems. However, concerns about the validity of these data often constrain their utility. In this paper, we focus on the use of citizen science data in addressing complex challenges in environmental conservation. We consider this issue from three perspectives. First, we present a literature scan of papers that have employed Bayesian models with citizen science in ecology. Second, we compare several popular majority vote algorithms and introduce a Bayesian item response model that estimates and accounts for participants' abilities after adjusting for the difficulty of the images they have classified. The model also enables participants to be clustered into groups based on ability. Third, we apply the model in a case study involving the classification of corals from underwater images from the Great Barrier Reef, Australia. We show that the model achieved superior results in general and, for difficult tasks, a weighted consensus method that uses only groups of experts and experienced participants produced better performance measures. Moreover, we found that participants learn as they have more classification opportunities, which substantially increases their abilities over time. Overall, the paper demonstrates the feasibility of CS for answering complex and challenging ecological questions when these data are appropriately analysed. This serves as motivation for future work to increase the efficacy and trustworthiness of this emerging source of data. [ABSTRACT FROM AUTHOR]

Of course, Rupert and Lincoln were part of the stat department, but in that department itself everything math ruled. B BN: b You have several papers on empirical Bayes with Carl Morris, Rob Tibshirani and even, you know, we worked on the empirical Bayes deconvolution stuff (Narasimhan & Efron, B., 2020). B BE: b Actually, one of the book-like things that I did that was most successful was the little monograph, when the bootstrap first came out, I The Jackknife, the Bootstrap, and other resampling plans i (Efron, 1982). Https://dx.doi.org/10.1017/CBO9781316576533 12 Efron, B., Hastie, T., Johnstone, I. & Tibshirani, R.J. 2004. Biometrika, 63, 435 - 447. http://www.jstor.org/stable/2335721 17 Efron, B. & Tibshirani, R.J. 1993. [Extracted from the article]

We would like to congratulate Professor Brad Efron for the well-deserved honor of winning the 2019 International Statistics Prize, and for his thought-provoking paper discussing the diverging paths of so-called traditional statistical regression and pure prediction approaches. In this discussion, we provide a portrayal of our own perspectives on the intersection of statistics and machine learning methods, and compare them with Professor Efron's. Professor Efron discusses this idea but dismisses these modern attribution attempts as being suboptimal, dampened by the pure prediction method's focus on prediction using many weak learners that do not prioritize individual strong predictors. [Extracted from the article]

Summary: Luis Raúl Pericchi Guerra was born in Caracas, Venezuela, on 11 March 1952. He completed a B.S. in Mathematics in 1975 at the Universidad Simón Bolívar in Caracas, an M.S. in Statistics at the University of California Berkeley in 1978 and a Ph.D. in Statistics at Imperial College London in 1981. After graduating from Imperial College, Luis Raúl went back to Universidad Simón Bolívar. There, he played a key role in the developing of graduate programmes in Statistics and single handedly built an internationally recognised group focused on Bayesian statistics. In 2001, he moved to the Universidad de Puerto Rico in Rio Piedras to become the Chair of the Mathematics Department. At Universidad de Puerto Rico, he was instrumental in the establishment of a Ph.D. track in Computational Mathematics and Statistics. Luis Raúl has published over 120 papers in statistical and domain‐specific journals, making significant contributions to several areas of Bayesian statistics (especially in the areas of model selection and Bayesian robustness) and their application (especially in hydrology). He is a Fellow of the American Statistical Association, the International Society for Bayesian Analysis, the John Simon Guggenheim Memorial Foundation and an Elected Member of the International Statistical Institute. This conversation took place over multiple sessions during the 2022 O'Bayes meeting in Santa Cruz, California, and the months that followed. [ABSTRACT FROM AUTHOR]

Summary: The advent of artificial intelligence (AI) and machine learning algorithms has led to opportunities as well as challenges in their use. In this overview paper, we begin with a discussion of bias and fairness issues that arise with the use of AI techniques, with a focus on supervised machine learning algorithms. We then describe the types and sources of data bias and discuss the nature of algorithmic unfairness. In addition, we provide a review of fairness metrics in the literature, discuss their limitations, and describe de‐biasing (or mitigation) techniques in the model life cycle. [ABSTRACT FROM AUTHOR]

Summary: The effective sample size (ESS) is widely used in sample‐based simulation methods for assessing the quality of a Monte Carlo approximation of a given distribution and of related integrals. In this paper, we revisit the approximation of the ESS in the specific context of importance sampling. The derivation of this approximation, that we will denote as ESS^, is partially available in a 1992 foundational technical report of Augustine Kong. This approximation has been widely used in the last 25 years due to its simplicity as a practical rule of thumb in a wide variety of importance sampling methods. However, we show that the multiple assumptions and approximations in the derivation of ESS^ make it difficult to be considered even as a reasonable approximation of the ESS. We extend the discussion of the ESS^ in the multiple importance sampling setting, we display numerical examples and we discuss several avenues for developing alternative metrics. This paper does not cover the use of ESS for Markov chain Monte Carlo algorithms. [ABSTRACT FROM AUTHOR]

Summary: This paper proposes an instrumental variable regression estimator of population totals using a sample, a set of links between the sample units and records on an administrative file, and a set of calibration totals calculated from the administrative file. This paper proposes a survey‐weighted estimator of a population total that is valid when the survey non‐response mechanism is non‐ignorable and false negatives occur in the administrative‐survey linkage. False negatives lead to measurement error in the administrative variables that are available on the survey and will lead to biased estimates if not taken into account. We show the benefit of the proposed approach in a simulation and in a case study. [ABSTRACT FROM AUTHOR]

Summary When using the customary direct questioning approach to collect sensitive information in a survey, some respondents may have no problem disclosing their true status or opinion, while others may be reluctant to reveal them. The problem of response bias is likely to arise as some responses tend to be more socially desirable than truthful. Hence, the indirect questioning approach can be used to guarantee privacy protection and reduce the influence of social desirability bias. This paper introduces a new technique which combines, both the direct and indirect questioning approaches and allows Bayesian estimates of the prevalence of multisensitive attributes to be obtained by taking into account the estimated proportion of honest respondents under direct questioning. Our proposal stems from a real survey in Taiwan and is illustrated with two motivating examples concerning voting behaviours and sexual identity. The empirical analysis reinforces that the proposed two‐stage multilevel method is satisfactory in mitigating the effects of respondents' self‐protective behaviour, and produces results that are more reliable than those based on the traditional direct questioning approach and better than a previous version of the multilevel randomised response method that ignores the presence of cheating behaviours in the direct questioning stage. [ABSTRACT FROM AUTHOR]

Summary: As sources of data become more plentiful and massive datasets are easier to acquire, new ethical issues arise involving data quality and privacy, and the analysis, interpretation and dissemination of data‐driven decisions. There are numerous anecdotes involving abuses of complex data analyses and algorithms, and the impact they have had on society. In this paper, we discuss what statisticians can do to help enhance data science ethics in practice and what statistics educators can do to instil sound ethical behaviour in our students. We have opportunities to practice and teach ethical conduct relevant to all stages of the data life cycle. This paper discusses issues impacting ethical data science, with a focus on how statisticians can help raise awareness and encourage implementation of ethical best practices. [ABSTRACT FROM AUTHOR]

Summary: Most modern articles in the palaeomagnetism literature are based on statistics developed by Fisher's 1953 paper 'Dispersion on a sphere', which assumes independent and identically distributed (iid) spherical data. However, palaeomagnetic sample designs are usually hierarchical, where specimens are collected within sites and the data are then combined across sites to calculate an overall mean direction for a geological formation. The specimens within sites are typically more similar than specimens between different sites, and so the iid assumptions fail. This article has three principal goals. The first is to review, contrast and compare both the statistics and geophysics literature on the topic of analysis methods for clustered data on spheres. The second is to present a new hierarchical parametric model, which avoids the unrealistic assumption of rotational symmetry in Fisher's 1953 paper 'Dispersion on a sphere' and may be broadly useful in the analysis of many palaeomagnetic datasets. To help develop the model, we use publicly available data as a case study collected from the Golan Heights volcanic plateau. The third goal is to explore different methods for constructing confidence regions for the overall mean direction based on clustered data. Two bootstrap confidence regions that we propose perform well and will be especially useful to geophysics practitioners. [ABSTRACT FROM AUTHOR]

A hot new statistical technology like the pure prediction algorithms divides the discipline into insiders and outsiders. The discussants here span the insider/outsider spectrum, including some of the most prominent developers of modern prediction methods. But, as Candes/Sabatti say, there are ongoing attempts to use the algorithms for causality purposes. [Extracted from the article]

Summary: Paul Embrechts was born in Schoten, Belgium, on 3 February 1953. He holds a Licentiaat in Mathematics from Universiteit Antwerpen (1975) and a DSc from Katholieke Universiteit Leuven (1979), where he was also a Research Assistant from 1975 to 1983. He then held a lectureship in Statistics at Imperial College, London (1983–1985) and was a Docent at Limburgs Universitair Centrum, Belgium (1985–1989) before joining ETH Zürich as a Full Professor of Mathematics in 1989, where he remained until his retirement as an Emeritus in 2018. A renowned specialist of extreme‐value theory and quantitative risk management, he authored or coauthored nearly 200 scientific papers and five books, including the highly influential 'Modelling of Extremal Events for Insurance and Finance' (Springer, 1997) and 'Quantitative Risk Management: Concepts, Techniques and Tools' (Princeton University Press, 2005, 2015). He served in numerous editorial capacities, notably as Editor‐in‐Chief of the ASTIN Bulletin (1996–2005). Praised for his natural leadership and exceptional communication skills, he helped to bridge the gap between academia and industry through the foundation of RiskLab Switzerland and his sustained leadership for nearly 20 years. He gave numerous prestigious invited and keynote lectures worldwide and served as a member of the board of, or consultant for, various banks, insurance companies and international regulatory authorities. His work was recognised through several visiting positions, including at the Oxford‐Man Institute, and many awards. He is, inter alia, an Elected Fellow of the Institute of Mathematical Statistics (1995) and the American Statistical Association (2014), an Honorary Fellow of the Institute and the Faculty of Actuaries (2000), Honorary Member of the Belgian (2010) and French (2015) Institute of Actuaries and was granted four honorary degrees (University of Waterloo, 2007; Heriot‐Watt University, 2011; Université catholique de Louvain, 2012; City, University of London, 2017). The following conversation took place in Paul's office at ETH Zürich, 17–18 December 2018. [ABSTRACT FROM AUTHOR]

Jim Ramsay was born on September 5, 1942, in Prince George, British Columbia. He pursued undergraduate studies at the University of Alberta, where he completed a BEd in 1964 with a major in English and a minor in mathematics. He then specialized in statistics and psychometry, earning a PhD in psychology from Princeton University in 1966. After holding a temporary lectureship in the Department of Psychology at University College London for one year, he joined the Department of Psychology at McGill University, where he rose through the academic ranks. He was chair of his department from 1986 to 1989 and spent sabbatical leaves in Cambridge, Grenoble, and Toulouse. He was named professor emeritus upon his retirement in 2007. Jim is the author of four influential books and over 100 peer-reviewed articles in statistical and psychometric journals. He developed much of the statistical theory behind multidimensional scaling and is widely recognized as the founder of functional data analysis. Three of his papers were read to the Royal Statistical Society, and another won The Canadian Journal of Statistics 2000 Best Paper Award. The Statistical Society of Canada (SSC) awarded him a Gold Medal for research in 1998 and an honorary membership in 2012. Jim was president of the Psychometric Society in 1981-82 and president of the SSC in 2002-03. The following conversation took place at Jim's home in Ottawa, Ontario, on March 14 and April 4, 2012. [ABSTRACT FROM AUTHOR]

Summary: Estimating the size of a hard‐to‐count population is a challenging matter. We consider uni‐list approaches in which the count of identifications per unit is the basis of analysis. Unseen units have a zero count and do not occur in the sample leading to a zero‐truncated setting. Because of various mechanisms, one‐inflation is often an occurring phenomena that can lead to seriously biased estimates of population size. The current work reviews some recent advances on one‐inflation and zero‐truncation modelling, and furthermore focuses here on the impact it has on population size estimation. The zero‐truncated one‐inflated and the one‐inflated zero‐truncated model is compared (also with the model ignoring one‐inflation) in terms of Horvitz–Thompson estimation of population size. The simulation work shows clearly the biasing effect of ignoring one‐inflation. Both models, the zero‐truncated one‐inflated and the one‐inflated zero‐truncated one, are suitable to model ongoing one‐inflation. It is also important to choose an appropriate base‐line distributional model. Finally, all models derived in the paper are illustrated on a number of case studies. [ABSTRACT FROM AUTHOR]

Summary: Since the pioneering work of sliced inverse regression, sufficient dimension reduction has been growing into a mature field in statistics and it has broad applications to regression diagnostics, data visualisation, image processing and machine learning. In this paper, we provide a review of several popular inverse regression methods, including sliced inverse regression (SIR) method and principal hessian directions (PHD) method. In addition, we adopt a conditional characteristic function approach and develop a new class of slicing‐free methods, which are parallel to the classical SIR and PHD, and are named weighted inverse regression ensemble (WIRE) and weighted PHD (WPHD), respectively. Relationship with recently developed martingale difference divergence matrix is also revealed. Numerical studies and a real data example show that the proposed slicing‐free alternatives have superior performance than SIR and PHD. [ABSTRACT FROM AUTHOR]

Summary Clustering of longitudinal data is becoming increasingly popular in many fields such as social sciences, business, environmental science, medicine and healthcare. However, it is often challenging due to the complex nature of the data, such as dependencies between observations collected over time, missingness, sparsity and non‐linearity, making it difficult to identify meaningful patterns and relationships among the data. Despite the increasingly common application of cluster analysis for longitudinal data, many existing methods are still less known to researchers, and limited guidance is provided in choosing between methods and software packages. In this paper, we review several commonly used methods for clustering longitudinal data. These methods are broadly classified into three categories, namely, model‐based approaches, algorithm‐based approaches and functional clustering approaches. We perform a comparison among these methods and their corresponding R software packages using real‐life datasets and simulated datasets under various conditions. Findings from the analyses and recommendations for using these approaches in practice are discussed. [ABSTRACT FROM AUTHOR]

Summary Fast online surveys without sampling frames are becoming increasingly important in survey research. Their recruitment methods result in non‐probability samples. As the mechanism of data generation is always unknown in such samples, the problem of non‐ignorability arises making vgeneralisation of calculated statistics to the population of interest highly questionable. Sensitivity analyses provide a valuable tool to deal with non‐ignorability. They capture the impact of different sample selection mechanisms on target statistics. In 2019, Andridge and colleagues proposed an index to quantify potential (non‐ignorable) selection bias in proportions that combines the effects of different selection mechanisms. In this paper, we validate this index with an artificial non‐probability sample generated from a large empirical data set and additionally applied it to proportions estimated from data on current political attitudes arising from a real non‐probability sample selected via River sampling. We find a number of conditions that must be met for the index to perform meaningfully. When these requirements are fulfilled, the index shows an overall good performance in both of our applications in detecting and correcting present selection bias in estimated proportions. Thus, it provides a powerful measure for evaluating the robustness of results obtained from non‐probability samples. [ABSTRACT FROM AUTHOR]

Summary Peter J. Rousseeuw is a statistician known mainly for his work on robust statistics and cluster analysis. Among his creations are least trimmed squares regression, the minimum covariance determinant estimator, the partitioning around medoids clustering method and the silhouettes graphical display. Peter obtained his PhD in 1981 following research carried out at the ETH in Zürich, Switzerland, which led to a book on influence functions. Later, he was a professor at Delft University of Technology, The Netherlands, and at the University of Antwerp, Belgium. Next, he was a researcher at Renaissance Technologies in New York for over a decade. He then returned to Belgium as a full professor at KU Leuven, until becoming emeritus in 2022. He is an elected member of the International Statistical Institute and a fellow of the Institute of Mathematical Statistics and the American Statistical Association. In the course of his career, Peter published three books and over 200 papers, together receiving over 100 000 citations. He was awarded the George Box Medal for Business and Industrial Statistics, the Research Medal of the International Federation of Classification Societies, the Frank Wilcoxon Prize, and twice the Jack Youden Prize. Recently, Peter received the 2024 ASA Noether Distinguished Scholar Award for nonparametric statistics. His former PhD students include Annick Leroy, Rik Lopuhaä, Geert Molenberghs, Christophe Croux, Mia Hubert, Stefan Van Aelst, Tim Verdonck and Jakob Raymaekers. He is the creator and sole sponsor of the Rousseeuw Prize for Statistics, which was first handed out by the King of Belgium in 2022. [ABSTRACT FROM AUTHOR]

Summary: Marc Hallin was born in Ghent, Belgium, on 23 April 1949. He holds a Licence en Sciences mathématiques (1971), a Licence en Sciences actuarielles (1972), and a Doctorat en Sciences (1976) from the Université libre de Bruxelles. He then rose through the professorial ranks at the same institution, being successively Premier Assistant (1977–1978), Chargé de Cours associé (1978–1984), Chargé de Cours (1984–1988), Professeur ordinaire (1988–2009), and Professeur ordinaire émérite upon retirement in 2009. Throughout his career, he supervised 25 PhD students and held invited positions at many institutions of high standing in Austria, Belgium, England, France, Hong Kong, Italy, Portugal, Spain, Switzerland, and the USA (most notably Princeton). A renown expert in time series analysis, econometrics, and non‐parametric inference, Marc is the author or coauthor of over 250 research papers, for which he received numerous awards, including the Medal of the Faculty of Mathematics and Physics of Charles University in Prague (2006), a Humboldt Forschungspreis from the Alexander von Humboldt Foundation (2012), the Pierre‐Simon de Laplace Award of the Société française de Statistique (2022), and the Gottfried E. Noether Distinguished Scholar Award of the American Statistical Association (2022). He gave several distinguished lecture series, including the 2017 Hermann Otto Hirschfeld Lecture Series at the Humboldt Universität zu Berlin, and the 2018 Mahalanobis Memorial Lecture at the Indian Statistical Institute. Over the years, he co‐edited a dozen books and proceedings, and served on the editorial boards of several journals, including the Journal of Time Series Analysis (1994–2009), the Journal of Econometrics (2013–2019), the Journal of Business and Economic Statistics (2018–), and the Theory and Methods Section of the Journal of the American Statistical Association (2005–). He is a Fellow of the Institute of Mathematical Statistics (1990) and the American Statistical Association (1997), as well as a member of the Classe des Sciences of the Royal Academy of Belgium (1999). Marc has been a member of the International Statistical Institute since 1985 and was (co‐) Editor‐in‐Chief of the International Statistical Review from 2010 to 2015. [ABSTRACT FROM AUTHOR]

Summary: In this paper, we survey the recent development of inversion‐free Newton's method, which directly avoids computing the inversion of Hessian, and demonstrate its applications in estimating parameters of models such as linear and logistic regression. A detailed review of existing methodology is provided, along with comparisons of various competing algorithms. We provide numerical examples that highlight some deficiencies of existing approaches, and demonstrate how the inversion‐free methods can improve performance. Motivated by recent works in literature, we provide a unified subsampling framework that can be combined with the inversion‐free Newton's method to estimate model parameters including those of linear and logistic regression. Numerical examples are provided for illustration. [ABSTRACT FROM AUTHOR]

Summary: Cointegration has shown useful for modeling non‐stationary data with long‐run equilibrium relationships among variables, with applications in many fields such as econometrics, climate research and biology. However, the analyses of vector autoregressive models are becoming more difficult as data sets of higher dimensions are becoming available, in particular because the number of parameters is quadratic in the number of variables. This leads to lack of statistical robustness, and regularisation methods are paramount for obtaining valid estimates. In the last decade, many papers have appeared suggesting different penalisation approaches to the inference problem. Here, we make a comprehensive review of different penalisation methods adapted to the specific structure of vector cointegrated models suggested in the literature, with relevant references to software packages. The methods are evaluated and compared according to a range of error measures in a simulation study, considering combinations of low and high dimension of the system and small and large sample sizes. [ABSTRACT FROM AUTHOR]

Summary: Graphical modelling is an important branch of statistics that has been successfully applied in biology, social science, causal inference and so on. Graphical models illuminate connections between many variables and can even describe complex data structures or noisy data. Graphical models have been combined with supervised learning techniques such as regression modelling and classification analysis with multi‐class responses. This paper first reviews some fundamental graphical modelling concepts, focusing on estimation methods and computational algorithms. Several advanced topics are then considered, delving into complex graphical structures and noisy data. Applications in regression and classification are considered throughout. [ABSTRACT FROM AUTHOR]

Summary: The integrated nested Laplace approximation (INLA) is a well‐known and popular technique for spatial modelling with a user‐friendly interface in the R‐INLA package. Unfortunately, only a certain class of latent Gaussian models are amenable to fitting with INLA. In this paper, we review template model builder (TMB), an existing technique and software package which is well‐suited to fitting complex spatio‐temporal models. TMB is relatively unknown to the spatial statistics community, but it is a flexible random effects modelling tool which allows users to define customizable and complex mixed effects models through C++ templates. After contrasting the methodology behind TMB with INLA, we provide a large‐scale simulation study assessing and comparing R‐INLA and TMB for continuous spatial models, fitted via the stochastic partial differential equations (SPDE) approximation. The results show that the predictive fields from both methods are comparable in most situations even though TMB estimates for fixed or random effects may have slightly larger bias than R‐INLA. We also present a smaller discrete spatial simulation study, in which both approaches perform well. We conclude with a joint analysis of breast cancer incidence and mortality data implemented in TMB which requires a model which cannot be fit with R‐INLA. [ABSTRACT FROM AUTHOR]

Summary: The fused lasso signal approximator (FLSA) is a smoothing procedure for noisy observations that uses fused lasso penalty on unobserved mean levels to find sparse signal blocks. Several path algorithms have been developed to obtain the whole solution path of the FLSA. However, it is known that the FLSA has model selection inconsistency when the underlying signals have a stair‐case block, where three consecutive signal blocks are either strictly increasing or decreasing. Modified path algorithms for the FLSA have been proposed to guarantee model selection consistency regardless of the stair‐case block. In this paper, we provide a comprehensive review of the path algorithms for the FLSA and prove the properties of the recently modified path algorithms' hitting times. Specifically, we reinterpret the modified path algorithm as the path algorithm for local FLSA problems and reveal the condition that the hitting time for the fusion of the modified path algorithm is not monotone in a tuning parameter. To recover the monotonicity of the solution path, we propose a pathwise adaptive FLSA having monotonicity with similar performance as the modified solution path algorithm. Finally, we apply the proposed method to the number of daily‐confirmed cases of COVID‐19 in Korea to identify the change points of its spread. [ABSTRACT FROM AUTHOR]

In 1903 the well-known Dutch astronomer Kapteyn published a paper in which he discussed statistical methods which he thought would be relevant for biologists. His motivation was the 1895 paper of Pearson on skew frequency curves. Kapteyn had concluded that the theory was open to grave objections and was not adapted to nonmathematical readers. He was then led to an independent investigation of the subject. This publication would lead to a heated dialogue between Kapteyn and Pearson, in which they accused each other of inappropriate starting points, of plagiarism and of making serious mathematical mistakes. This article evaluates these claims. In addition, Kapteyn's struggle to make his work accessible to biologists is discussed. [ABSTRACT FROM AUTHOR]

Summary: This article provides an overview of supervised machine learning (ML) with a focus on applications in banking. The supervised ML techniques covered include bagging (random forest), boosting (gradient boosting machine) and neural networks. We begin with an introduction to ML tasks and techniques. This is followed by a description of tree‐based ensemble algorithms, including bagging with random forest and boosting with gradient boosting machines, as well as feedforward neural networks. We then provide an extensive discussion of hyper‐parameter optimisation techniques. Interpretability of ML results is an important topic in banking and other regulated industries, and it is also covered in some depth. The paper concludes with a comparison of the features of different ML algorithms and a discussion of their use in practice. An application from credit risk modelling in banking is used throughout the paper to illustrate the techniques and interpret the results of the algorithms. [ABSTRACT FROM AUTHOR]

Summary: Finding a suitable representation of multivariate data is fundamental in many scientific disciplines. Projection pursuit (PP) aims to extract interesting 'non‐Gaussian' features from multivariate data, and tends to be computationally intensive even when applied to data of low dimension. In high‐dimensional settings, a recent work (Bickel et al., 2018) on PP addresses asymptotic characterization and conjectures of the feasible projections as the dimension grows with sample size. To gain practical utility of and learn theoretical insights into PP in an integral way, data analytic tools needed to evaluate the behaviour of PP in high dimensions become increasingly desirable but are less explored in the literature. This paper focuses on developing computationally fast and effective approaches central to finite sample studies for (i) visualizing the feasibility of PP in extracting features from high‐dimensional data, as compared with alternative methods like PCA and ICA, and (ii) assessing the plausibility of PP in cases where asymptotic studies are lacking or unavailable, with the goal of better understanding the practicality, limitation and challenge of PP in the analysis of large data sets. [ABSTRACT FROM AUTHOR]

Summary: In this paper, we perform a sparse filtering recursion for efficient changepoint detection for discrete‐time observations. We attach auxiliary event times to the chronologically ordered observations and formulate multiple changepoint problems of discrete‐time observations into continuous‐time observations. Ideally, both the computational and memory costs of the proposed auxiliary uniformisation forward‐filtering backward‐sampling algorithm can be quadratically scaled down to the number of changepoints instead of the number of observations, which would otherwise be prohibitive for a long sequence of observations. To avoid model bias, a time‐varying changepoint recurrence rate across different segments is assumed to characterise diverse scales of run lengths of the changepoints. We demonstrate the methods through simulation studies and real data analysis. [ABSTRACT FROM AUTHOR]

Summary: Power laws and power laws with exponential cut‐off are two distinct families of distributions on the positive real half‐line. In the present paper, we propose a unified treatment of both families by building a family of distributions that interpolates between them, which we call Interpolating Family (IF) of distributions. Our original construction, which relies on techniques from statistical physics, provides a connection for hitherto unrelated distributions like the Pareto and Weibull distributions, and sheds new light on them. The IF also contains several distributions that are neither of power law nor of power law with exponential cut‐off type. We calculate quantile‐based properties, moments and modes for the IF. This allows us to review known properties of famous distributions on ℝ+ and to provide in a single sweep these characteristics for various less known (and new) special cases of our Interpolating Family. [ABSTRACT FROM AUTHOR]

Summary: This paper reviews methods for handling complex sampling schemes when analysing categorical survey data. It is generally assumed that the complex sampling scheme does not affect the specification of the parameters of interest, only the methodology for making inference about these parameters. The organisation of the paper is loosely chronological. Contingency table data are emphasised first before moving on to the analysis of unit‐level data. Weighted least squares methods, introduced in the mid 1970s along with methods for two‐way tables, receive early attention. They are followed by more general methods based on maximum likelihood, particularly pseudo maximum likelihood estimation. Point estimation methods typically involve the use of survey weights in some way. Variance estimation methods are described in broad terms. There is a particular emphasis on methods of testing. The main modelling methods considered are log‐linear models, logit models, generalised linear models and latent variable models. There is no coverage of multilevel models. [ABSTRACT FROM AUTHOR]

Nan Laird has an enormous and growing impact on computational statistics. Her paper with Dempster and Rubin on the expectation-maximisation (EM) algorithm is the second most cited paper in statistics. Her papers and book on longitudinal modelling are nearly as impressive. In this brief survey, we revisit the derivation of some of her most useful algorithms from the perspective of the minorisation-maximisation (MM) principle. The MM principle generalises the EM principle and frees it from the shackles of missing data and conditional expectations. Instead, the focus shifts to the construction of surrogate functions via standard mathematical inequalities. The MM principle can deliver a classical EM algorithm with less fuss or an entirely new algorithm with a faster rate of convergence. In any case, the MM principle enriches our understanding of the EM principle and suggests new algorithms of considerable potential in high-dimensional settings where standard algorithms such as Newton's method and Fisher scoring falter.

Summary A valuable and robust extension of the traditional joint mean and the covariance models when data subject to outliers and/or heavy‐tailed outcomes can be achieved using the joint modelling of location and scatter matrix of the multivariate t‐distribution. This model encompasses three models in itself, and the number of unknown parameters in the covariance model increases quadratically with the matrix size. As a result, selecting the important variables becomes a crucial aspect to consider. In this context, the variable selection combined with the parameter estimation is considered under the normality assumption. However, because of the non‐robustness of the normal distribution, the resulting estimators will be sensitive to outliers and/or heavy taildness in the data. This paper has two objectives to overcome these problems. The first is to obtain the maximum likelihood estimates of the parameters and propose an expectation‐maximisation type algorithm as an alternative to the Fisher scoring algorithm in the literature. We also consider simultaneous parameter estimation and variable selection in the multivariate t‐joint location and scatter matrix models. The consistency and oracle properties of the regularised estimators are also established. Simulation studies and real data analysis are provided to assess the performance of the proposed methods. [ABSTRACT FROM AUTHOR]

Summary Multicollinearity results in inflation in the variance of the ordinary least squares estimators due to the correlation between two or more independent variables (including the constant term). A widely applied solution is to estimate with penalised estimators such as the ridge estimator, which trade off some bias in the estimators to gain a reduction in the variance of these estimators. Although the variance diminishes with these procedures, all seem to indicate that the inference and goodness of fit are controversial. Alternatively, the raise regression allows mitigation of the problems associated with multicollinearity without the loss of inference or the coefficient of determination. This paper completely formalises the raise estimator. For the first time, the norm of the estimator, the behaviour of the individual and joint significance, the behaviour of the mean squared error and the coefficient of variation are analysed. We also present the generalisation of the estimation and the relation between the raise and the residualisation estimators. To have a better understanding of raise regression, previous contributions are also summarised: its mean squared error, the variance inflation factor, the condition number, adequate selection of the variable to be raised, the successive raising, and the relation between the raise and the ridge estimator. The usefulness of the raise regression as an alternative to mitigate multicollinearity is illustrated with two empirical applications. [ABSTRACT FROM AUTHOR]

Summary: In this paper, we develop a fully parametric quantile regression model based on the generalised three‐parameter beta (GB3) distribution. Beta regression models are primarily used to model rates and proportions. However, these models are usually specified in terms of a conditional mean. Therefore, they may be inadequate if the observed response variable follows an asymmetrical distribution. In addition, beta regression models do not consider the effect of the covariates across the spectrum of the dependent variable, which is possible through the conditional quantile approach. In order to introduce the proposed GB3 regression model, we first reparameterise the GB3 distribution by inserting a quantile parameter, and then we develop the new proposed quantile model. We also propose a simple interpretation of the predictor–response relationship in terms of percentage increases/decreases of the quantile. A Monte Carlo study is carried out for evaluating the performance of the maximum likelihood estimates and the choice of the link functions. Finally, a real COVID‐19 dataset from Chile is analysed and discussed to illustrate the proposed approach. [ABSTRACT FROM AUTHOR]

Summary: This paper proposes a deep generalisation for income inequality indices. A generalised income inequality index that depends on two parameters and that involves a large set of income inequality indices in the same framework is proposed. The two parameters control the sensitivity of the generalised index to different levels of the income distribution. A thorough investigation of the generalised index paves the way for understanding the influence of the low, middle and high incomes on various income inequality indices and thereby facilitates the choice of multiple indices simultaneously for a better analysis of inequality as advocated by several recent studies. Moreover, two methods for estimating the generalised index in the case of finite populations are shown. A new method for estimating the inequality indices is proposed. [ABSTRACT FROM AUTHOR]