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.     

The Case Study Method in Psychodynamic Psychology: Focus on Addiction

The case study method has been essential in psychoanalysis and psychodynamic therapy, since it is the only way to describe and explore the deepest levels of the human psyche. Addiction is no more and no less than a particular psychological mechanism, identical at its core to other psychological compulsions, and is therefore best understood and reported by this method that explores the mind in depth. We will discuss the value of the case report method in general and in specific with regard to psychoanalysis and addiction, criticisms raised about this method, and comparisons of it with nomothetic research.     

‘By Helping Others,We Helped Ourselves:’ Volunteering and Social Integration of Ex-Soviet Immigrants in Israel

Recent immigrants seldom join the ranks of volunteers for various social causes. Immigrants from former socialist countries have been shown to be particularly averse to organized forms of volunteering for reasons rooted in their past, including forced forms of collectivism imposed by the state. In this qualitative study, we explored the perceptions and practices of volunteering among ex-Soviet immigrants (mostly educated middle-aged women) who ran a project for the benefit of elderly. Our findings show that most volunteers chose causes targeting fellow immigrants, their resettlement and well-being, and were motivated by the wish to build co-ethnic support network and overcome marginalization in the Israeli society. Other volunteers were driven by the need for self-actualization in the context of underemployment and occupational downgrading. Personal empowerment and higher identification with the receiving society were the most salient outcomes of volunteering for our informants. We conclude that for some immigrants, volunteering can serve as a strategy of social integration.     

The Nigerian Civil Registration and Vital Statistics System: Contexts,Institutions, Operation

Civil Registration and Vital Statistics (CRVS) Systems are playing increasing roles in national and global data infrastructure, facilitating the achievement of key developmental goals. The performance of CRVS systems in many developing countries including Nigeria is sub optimal and information on their structure and operations scanty. This paper characterizes the statutory, institutional, and operational elements and social contexts of the Nigerian CRVS System. It includes a context analysis of the system and a review of improvement strategies and interventions. The paper shows that the Nigerian CRVS System is built on viable statutory and institutional platforms—the constitutional enshrinement of: universal, compulsory and continuous civil registration and, a central registration institution—the National Population Commission (NPC). NPC has registration hierarchy and infrastructure deployed at every level of political administration and mechanisms for efficient coordination of registration and production of vital statistics on live-births, deaths and stillbirths amongst other vital events. The major weaknesses of the system include inadequate financing and deployment of registration facilities, conflicting statutes and institutional arrangements for the registration of key events like death and stillbirths. Other weaknesses include intra-organizational conflicts in resource allocation and weak ICT infrastructure within the NPC. These are compounded by overarching contextual issues namely registration unfriendly sociocultural norms, weak national data infrastructure, systemic corruption, poverty and undeveloped civic culture. Domestic initiatives and interventions to improve the system have focused on birth registration, while multilateral interventions have only made a modest impact on the system. Thus, holistic in-country improvement initiatives supported with robust interventions are imperative to overcome weaknesses in the Nigerian CRVS system. However fundamental improvements in CRVS systems in developing countries like Nigeria may likely only follow sustainable solutions to overarching contextual issues especially poor infrastructure, weak institutions, poverty and systemic corruption.     

Approximate Bayesian inference in semi-mechanistic models

Inference of interaction networks represented by systems of differential equations is a challenging problem in many scientific disciplines. In the present article, we follow a semi-mechanistic modelling approach based on gradient matching. We investigate the extent to which key factors, including the kinetic model, statistical formulation and numerical methods, impact upon performance at network reconstruction. We emphasize general lessons for computational statisticians when faced with the challenge of model selection, and we assess the accuracy of various alternative paradigms, including recent widely applicable information criteria and different numerical procedures for approximating Bayes factors. We conduct the comparative evaluation with a novel inferential pipeline that systematically disambiguates confounding factors via an ANOVA scheme.     

Maternal union instability and childhood mortality risk in the Global South, 2010–14

Efforts to improve child survival in lower-income countries typically focus on fundamental factors such as economic resources and infrastructure provision, even though research from post-industrial countries confirms that family instability has important health consequences. We tested the association between maternal union instability and children’s mortality risk in Africa, Latin America and the Caribbean, and Asia using children’s actual experience of mortality (discrete-time probit hazard models) as well as their experience of untreated morbidity (probit regression). Children of divorced/separated mothers experience compromised survival chances, but children of mothers who have never been in a union generally do not. Among children of partnered women, those whose mothers have experienced prior union transitions have a higher mortality risk. Targeting children of mothers who have experienced union instability—regardless of current union status—may augment ongoing efforts to reduce childhood mortality, especially in Africa and Latin America where union transitions are common.     

Heterogeneous climate effects on human migration in Indonesia

We examine the effect of anomalous temperatures, rainfall levels, and monsoon timing on migration outcomes in Indonesia. Using panel data from the Indonesian Family Life Survey and high-resolution climate data, we assess whether intra- and inter-province moves are used as a response to climatic shocks. We evaluate the relative importance of temperature, rainfall, and monsoon timing for migration. Only temperature and monsoon timing have significant effects, and these do not operate in the direction commonly assumed. Estimated effects vary according to individuals’ gender, membership in a farm household, and location. We also analyze climate effects on sources of household income, which highlights the multi-phasic nature of household responses. Results undermine narratives of a uniform global migratory response to climate change and highlight the heterogeneous use of migration as a response to such changes. By extending previous research on environmentally induced migration in Indonesia, we also highlight the sensitivity of estimates to alternative climate and migration measures.     

Effects of the social dominance orientation on environmental attitudes of Chilean university students / Efectos de la orientación a la dominancia social sobre las actitudes ambientales de universitarios chilenos

By adopting a multidimensional approach to environmental attitudes (apathy, anthropocentrism, connectedness and emotional affinity with nature), this study explores the relationships between these dimensions, social dominance orientation (SDO) and political orientation. Two hundred and sixty-two Chilean university students filled out an ad hoc online questionnaire. Two confirmatory factorial analyses showed the four typologies of environmental attitudes and the two factors for SDO, which are labelled group dominance and opposition to equality. Through hierarchical regressions, group dominance showed higher explanatory power of environmental attitudes than political orientation. Furthermore, using the bootstrap procedure we showed that group dominance mediated between political orientation and attitudes of apathy, anthropocentrism and connectedness, while opposition to equality mediated between apathy and emotional affinity. We conclude that SDO is an important variable for measuring political ideology while also providing new nuances when analysing the relationships with environmental attitudes.     

Disruptive Organisation

Organizations are driven by continuous change. Global, socio-economic and technological megatrends – demographic changes and digitalization – are the reasons for transformation. Traditional organization forms are not in line with customers‘ need. Disruptive modes of development and innovations have been intensively discussed elsewhere. The topic of disruptive organization have not been discussed towards internal and external customer relations, although of high relevance. 129 Managers answered on communication ways and the purpose of upcoming meaning for the working place/office. Two new models and five concrete recommendations have been established for upcoming future organization.     

Estimation and diagnostic analysis in skew-generalized-normal regression models

The skew-generalized-normal distribution [Arellano-Valle, RB, Gómez, HW, Quintana, FA. A new class of skew-normal distributions. Comm Statist Theory Methods 2004;33(7):1465–1480] is a class of asymmetric normal distributions, which contains the normal and skew-normal distributions as special cases. The main virtues of this distribution is that it is easy to simulate from and it also supplies a genuine expectation–maximization (EM) algorithm for maximum likelihood estimation. In this paper, we extend the EM algorithm for linear regression models assuming skew-generalized-normal random errors and we develop a diagnostics analyses via local influence and generalized leverage, following Zhu and Lee's approach. This is because Cook's well-known approach would be more complicated to use to obtain measures of local influence. Finally, results obtained for a real data set are reported, illustrating the usefulness of the proposed method.