Correlation and variable importance in random forests

This paper is about variable selection with the random forests algorithm in presence of correlated predictors. In high-dimensional regression or classification frameworks, variable selection is a difficult task, that becomes even more challenging in the presence of highly correlated predictors. Firstly we provide a theoretical study of the permutation importance measure for an additive regression model. This allows us to describe how the correlation between predictors impacts the permutation importance. Our results motivate the use of the recursive feature elimination (RFE) algorithm for variable selection in this context. This algorithm recursively eliminates the variables using permutation importance measure as a ranking criterion. Next various simulation experiments illustrate the efficiency of the RFE algorithm for selecting a small number of variables together with a good prediction error. Finally, this selection algorithm is tested on the Landsat Satellite data from the UCI Machine Learning Repository.     

How Does the Individual Find a Place in French Sociology?

In France over the last four decades increasing interest in the small, the specific and the subjective can be observed. This involves a shift away from a holist approach towards social analysis focused on the individual and often carried out by researchers as specified and reflexive individuals. To be understood, this movement must be situated in the context of the development of French sociology, largely dominated at its origins by Durkheimian sociology. Thus, when the individual has appeared this is often not in his or her own right, but as representative of a social trend or group. As from the 1980s, new perspectives, very often imported from American sociology bring into play individual action, emotions and bodies, often studied as socially shaped. More recently, individual characteristics have been examined as a constraint on the possibilities of action, and discussion has also revolved around the reality of the person as an autonomous subject.     

What is DIstinctive about Mediation? Experience and Expertise in Media

If great material resources and specialist technical knowledge are no longer required for individuals to act as mass communicators, what resources are required? This paper reviews social theories of production that can help to shed light on the matter. It explores the nature of knowledge required in the production of content meant for a mass audience, and the manner of its validation. The paper progressively conceptualizes mass communication or mediation in terms of ‘media experience and expertise.’ It begins by framing the discussion of media production in terms of mediation and symbolic/media power, and teasing out knowledge and competence as problems that merit special attention. It then draws on two parallel areas of scholarship to think through the blurring of formerly neat boundaries between interpersonal and mass communication; producer and audience; and expert and layperson. First, the renewed interest in the lived experience of professional producers that is spurred by media production ethnographers. Second, ongoing debates in Science and Technology Studies (STS) around knowledge production and public participation. The paper concludes with a discussion of outstanding issues with respect to media theory and proposals for future work.     

Impact of Gender Binarism on Hijras’ Life Course and Their Access to Fundamental Human Rights in Pakistan

This study adds to the growing body of knowledge on gender nonconformity aspects of heteronormativity by examining its impact on the life course of hijras and their access to fundamental human rights in Pakistan. Drawing on 50 semistructured interviews conducted in two sites, the findings suggest that the participants’ lived experiences associated with gender nonconformity significantly influenced the direction of their life course and their ability to have access to human rights. These experiences spanned from childhood to elderhood across a wide range of settings, such as family, school, guru dera (residence headed by a hijra guru), workplace, and interactions with authorities. The participants’ human rights were not recognized, resulting in abuse, social stigma, discrimination against them, and their exclusion from mainstream society. Finally, implications are drawn for public policy and future research on third gender concerns in Pakistan and elsewhere.     

Perception and Communication of Flood Risks: A Systematic Review of Empirical Research

Flood hazards are the most common and destructive of all natural disasters. For decades, experts have been examining how flood losses can be mitigated. Just as in other risk domains, the study of risk perception and risk communication has gained increasing interest in flood risk management. Because of this research growth, a review of the state of the art in this domain is believed necessary. The review comprises 57 empirically based peer‐reviewed articles on flood risk perception and communication from the Web of Science and Scopus databases. The characteristics of these articles are listed in a comprehensive table, presenting research design, research variables, and key findings. From this review, it follows that the majority of studies are of exploratory nature and have not applied any of the theoretical frameworks that are available in social science research. Consequently, a methodological standardization in measuring and analyzing people's flood risk perceptions and their adaptive behaviors is hardly present. This heterogeneity leads to difficulties in comparing results among studies. It is also shown that theoretical and empirical studies on flood risk communication are nearly nonexistent. The article concludes with a summary on methodological issues in the fields of flood‐risk perception and flood‐risk communication and proposes an agenda for future research.     

THE DEMAND FOR LIQUID ASSETS,CORPORATE SAVING,AND INTERNATIONAL CAPITAL FLOWS

The recent period of capital outflows from emerging economies has coincided with an increase in their corporate saving. In this paper, we model corporate saving as a demand for liquid assets by credit‐constrained firms in a dynamic open‐economy macroeconomic model. We find that the implications of this model are very different from standard models, because the demand for foreign bonds is a complement to domestic investment rather than a substitute. We show that this complementarity is at work when an emerging economy is on its convergence path or when it has a higher TFP growth rate. This framework is consistent with a number of stylized facts found in high‐growth, high‐investment emerging economies.     

An International Pooled Analysis for Obtaining a Benchmark Dose for Environmental Lead Exposure in Children

Lead is a recognized neurotoxicant, but estimating effects at the lowest measurable levels is difficult. An international pooled analysis of data from seven cohort studies reported an inverse and supra‐linear relationship between blood lead concentrations and IQ scores in children. The lack of a clear threshold presents a challenge to the identification of an acceptable level of exposure. The benchmark dose (BMD) is defined as the dose that leads to a specific known loss. As an alternative to elusive thresholds, the BMD is being used increasingly by regulatory authorities. Using the pooled data, this article presents BMD results and applies different statistical techniques in the analysis of multistudy data. The calculations showed only a limited variation between studies in the steepness of the dose‐response functions. BMD results were quite robust to modeling assumptions with the best fitting models yielding lower confidence limits (BMDLs) of about 0.1–1.0 μ g/dL for the dose leading to a loss of one IQ point. We conclude that current allowable blood lead concentrations need to be lowered and further prevention efforts are needed to protect children from lead toxicity.     

Mean estimation with data missing at random for functional covariables

In a missing-data setting, we want to estimate the mean of a scalar outcome, based on a sample in which an explanatory variable is observed for every subject while responses are missing by happenstance for some of them. We consider two kinds of estimates of the mean response when the explanatory variable is functional. One is based on the average of the predicted values and the second one is a functional adaptation of the Horvitz–Thompson estimator. We show that the infinite dimensionality of the problem does not affect the rates of convergence by stating that the estimates are root-n consistent, under missing at random (MAR) assumption. These asymptotic features are completed by simulated experiments illustrating the easiness of implementation and the good behaviour on finite sample sizes of the method. This is the first paper emphasizing that the insensitiveness of averaged estimates, well known in multivariate non-parametric statistics, remains true for an infinite-dimensional covariable. In this sense, this work opens the way for various other results of this kind in functional data analysis.     

A further examination of antecedents of correctional staff life satisfaction

Working in corrections is not only a demanding job, but a socially important one. While a growing number of studies have examined how the work environment impacts the job stress, job satisfaction, and organizational commitment of staff, very few studies have examined how working in corrections impacts the life satisfaction of workers. The current study utilized OLS regression to examine the antecedents of life satisfaction among staff at a Midwestern private prison. Job satisfaction had a positive relationship with life satisfaction, while age, work on family conflict, family on work conflict, and job involvement all had statistically significant negative effects. Finally, perceptions of the level of financial rewards, job stress, organizational commitment, gender, race, educational level, tenure, supervisory status, position, marital status, and having children, all had non-significant associations with overall satisfaction with life.