Search

Spatial data are becoming increasingly accessible to urban scientists, but these data are often prone to measurement error. Motivated by the analysis of the Milan (Italy) apartment market heterogeneity, we propose a semiparametric approach to adjust for the presence of measurement error in the covariates when estimating M-quantile regression. The M-quantile approach helps explain the heterogeneity across individual units, preserving robustness and efficiency in the estimates. The model’s parameters are estimated within a penalised likelihood framework and an analytical expression is proposed to estimate standard errors. Asymptotic properties of estimates are also provided.

This paper proposes an M-quantile regression approach to address the heterogeneity of the housing market in a modern European city. We show how M-quantile modelling is a rich and flexible tool for empirical market price data analysis, allowing us to obtain a robust estimation of the hedonic price function whilst accounting for different sources of heterogeneity in market prices. The suggested methodology can generally be used to analyse nonlinear interactions between prices and predictors. In particular, we develop a spatial semiparametric M-quantile model to capture both the potential nonlinear effects of the cultural environment on pricing and spatial trends. In both cases, nonlinearity is introduced into the model using appropriate bases functions. We show how the implicit price associated with the variable that measures cultural amenities can be determined in this semiparametric framework. Our findings show that the effect of several housing attributes and urban amenities differs significantly across the response distribution, suggesting that buyers of lower-priced properties behave differently than buyers of higher-priced properties.

In many applications of small area estimation, dichotomous or categorical outcomes are the targets of statistical inference. For example, in the analysis of labour markets, proportions of working-age people in the various labour market statuses are of interest. In this paper, in line with the recent literature, we consider a classification with more than three statuses and estimate related population parameters for 611 local labour market areas using data from the 2012 Italian Labour Force Survey, administrative registers and the 2011 Census. As for the methodology, we propose multinomial expectile regression models. These models provide a means to utilise (Formula presented.) -quantile type approaches, which have been shown to be a useful alternative to mixed model approaches when parametric assumptions on the distribution of random effects cannot be met. Via a large-scale simulation study, we show how this novel approach is much faster and provides reliable results when compared to multinomial mixed model approaches, and works for any number of categories rather than just a small number of categories as is more commonly the case with existing methods. Furthermore, the proposed approach potentially provides a framework for developing other methods for prediction with multi-category outcomes.

We consider maximum likelihood estimation of the Latent Class (LC) model, which is formulated through individual discrete latent variables. We explore tempering techniques to overcome the problem of multimodality of the log-likelihood function. A Tempered Expectation-Maximization (T-EM) algorithm is proposed, which can adequately explore the parameter space and reach the global maximum more frequently than the standard EM algorithm. We assess the performance of the proposed approach by a Monte Carlo simulation study and an application based on data about anxiety and depression in oncological patients.

We propose a hidden Markov model for longitudinal multivariate continuous responses, accounting for missing data under the missing at random assumption. Maximum likelihood estimation of this model is carried out through the Expectation-Maximization algorithm. To address the problem of dimensionality reduction, we develop a greedy search algorithm based on the Bayesian Information Criterion. We illustrate the proposal through a dataset collected by the World Bank and UNESCO Institute for Statistics on the basis of which we dynamically cluster countries according to the selected variables observed during the period 2000-2017.

In 2019, the European Union registered its seventh consecutive year of economic growth, but significant differences among the European Union countries are still present. In this context, competitiveness is an important element of human-centric and sustainable economic progress. Focusing on this concept, this study proposes an approach to track the dynamics of competitive growth in a set of member countries. After selecting some economic indicators for a 6-years period from 2014 to 2019, a time trajectory showing the dynamics of innovation and digitalization for each country has been plotted. This graphical analysis allows to extract some evidences about different paths to competitiveness within the European Union.

As recently stated by EUROSTAT, Small and Medium Enterprises (SMEs) play a crucial role in the economy of European Union (EU). This is particularly true for Italy that has one of the largest share of SMEs in the Euro zone. In such a context, the assessment of the Italian firms creditworthiness is a priority, especially in the early stage of a SMEs life. The aim of this work is to explore the effectiveness of the Feedforward Neural Network for Italian firms defaults prediction in comparison with standard linear models (Logistic and Probit). Our analysis considers the Italian manufacturing SMEs in a short time span (2016-2017). Results, obtained via two different sampling strategies, are not clear-cut.

The objective of credit scoring is to develop accurate rule of classification that aids to distinguish between good and bad clients. In this context, also Statistical Learning (SL) techniques have been explored, for building models that estimate the clients’ probability of insolvency. Although there are some encouraging results in literature, two main issues makes this classification task very hard: (i) high imbalance ratio between the two groups in the target variable and (ii) the effect of hyperparameter settings on overall performance. In this work, Bayesian Optimization (BO) is used to optimize the hyperparameters of a cost sensitive eXtreme Gradient Boosting (XGBoost) model. Experimental results reveal that the proposed solution is a promising starting point for future development

Questo lavoro studia la stratificazione sociale dei lavoratori stranieri in Italia con una prospettiva centrata sulla classe sociale: è così possibile analizzare la trasmissione delle disuguaglianze e la mobilità sociale (sia inter- che intragenerazionale) dal paese di origine a quello di destinazione. Si ipotizza che bassa origine sociale e posizione di classe nel paese d’origine siano associate a scarsi esiti occupazionali (Hyp 1) e a ridotte possibilità di mobilità ascendente (Hyp 2) nel paese di destinazione. I risultati sottolineano una chiara riproduzione di classe negli strati più bassi della struttura occupazionale. Inoltre, la mobilità occupazionale verso l’upper class (ma non verso la piccola borghesia) è più realizzabile per chi ha già un’elevata posizione sociale, mentre chi ha un basso status socio-economico incontra difficoltà a migliorare la propria posizione anche nel paese ospitante., In this work we aim at studying the social stratification of foreign workers in Italy adopting a class-based approach: this perspective allows us to shed light on the transmission of inequality and social mobility (both inter- and intra-generational) from country of origin to destination. We formulate the hypotheses that a low social origin and class position in home country is associated both with poor occupational achievement (Hyp 1) and with scarce chances of upward social mobility (Hyp 2) in the host country. Our results underline a clear class reproduction in the lowest strata of the occupational structure. Furthermore, occupational mobility to upper class (but not to petty bourgeoisie) is easier for who already has a high social position, while who lies at the bottom of the social structure is prevented from improving his\her position also in the host country.

Optimal individualized treatment estimation in single or multi-stage clinical trials is a breakthrough for personalized medicine. In this context, statistical methodologies can significantly improve the estimation over model-based methods. Furthermore it can help in selecting important variables in high-dimensional settings. In this work, we investigate the performance of an hybrid approach that combine Sequential Advantage Selection (SAS) method and Q-learning. In addition, Q-functions are estimated with linear regression model, random forest and neural network.

Clustering methods have typically found their application when dealing with continuous data. However, in many modern applications data consist of multiple categorical variables with no natural ordering. In the heuristic framework the problem of clustering these data is tackled by introducing suitable distances. In this work, we develop a model-based approach for clustering categorical data with nominal scale. Our approach is based on a mixture of distributions defined via the Hamming distance between categorical vectors. Maximum likelihood inference is delivered through an expectation-maximization algorithm. A simulation study is carried out to illustrate the proposed approach.

We propose a method to test dependence or conditional dependence be- tween parts of the domain of functional data. The tests are based on permutation procedure that tests if suitable blocks of the covariance or precision matrix of ba- sis expansion coefficients are equal to zero. We show that the procedure is able to identify the true structure of conditional dependence.

Nella stima robusta di un cluster weighted model, lo statistico deve fare molte scelte: specificare la forma dei cluster nelle variabili esplicative, assumere (o meno) varianza uguale per gli errori nelle linee di regressione e impostare i valori degli iper-parametri per la stima robusta, per evitare la distorsione generata da valori anomali e contaminazione. L’iper-parametro più delicato da specificare è la percentuale di trimming, ovvero la quantità di dati da escludere nella stima per garantirne l’affidabilità. In questo lavoro introduciamo specifici strumenti diagnostici per aiutare il professionista, o lo scienziato che ha bisogno di classificare i dati, a compiere una scelta ragionata a riguardo di tale iper-parametro, anche in base ad una prima esplorazione dello spazio delle soluzioni., In a robust approach to model fitting for the cluster weighted model, many choices are to be made by the statistician: specifying the shape of the clusters in the explanatory variables, assuming (or not) equal variance for the errors in the regression lines, and setting hyper-parameter values for the robust estimation to be protected from outliers and contamination. The most delicate hyper-parameter to specify is perhaps the percentage of trimming, or the amount of data to be excluded from the estimate, to ensure reliable inference. In this work we introduce diagnostic tools to help the professional, or the scientist who needs to group the data, to make an educated choice about this hyper-parameter, after a first exploration of the resulting model space.

Using the ‘Social Condition and Integration of Foreign Citizen’ survey (2011-2012) for Italy and the ‘Trajectoires et Origines’ survey (2008-2009) for France, we examined differences in the need for informal childcare and parental preferences in childcare (grandparents, other family and extra-family members) between migrants in Italy and France according to family demographic and socio-economic characteristics, migrants’ background and contextual factors. Results suggest that migrants’ choices and preferences are strongly affected by the household composition, mother occupational status and migrants’ country of origin. Furthermore, our results highlight the role played by the destination country, which can shape migrants’ choices and options. The use of informal childcare is higher for migrants in Italy than in France, even among the same origin areas.

In observational studies evaluating the treatment effect on a given out- come, the treated and untreated subjects may be highly unbalanced in their observed covariates, and these differences can lead to biased estimates of treatment effects. Propensity score is popular tool to reduce this bias. In this work we propose to esti- mate the propensity score by using Bayesian Networks as alternative to conventional logistic regression. Based on it, we develop an inferential methodology to evaluate the treatment effect. In simulation study, our proposed approach resulted in the best performance.

Data linkage can be used to combine values of the variable of interest from a national survey with values of auxiliary variables obtained from another source, such as a population register, for use in small area estimation. However, linkage errors can induce bias when fitting regression models; moreover, they can create non-representative outliers in the linked data in addition to the presence of potential representative outliers. In this paper, we adopt a secondary analyst’s point of view, assuming that limited information is available on the linkage process, and develop small area estimators based on linear mixed models and M-quantile models to accommodate linked data containing a mix of both types of outliers. We illustrate the properties of these small area estimators, as well as estimators of their mean squared error, by means of model-based and design-based simulation experiments. We further illustrate the proposed methodology by applying it to linked data from the European Survey on Income and Living Conditions and the Italian integrated archive of economic and demographic micro data in order to obtain estimates of the average equivalised income for labour market areas in central Italy.

In a standard classification framework, a discriminating rule is usually built from a trustworthy set of labeled units. In this context, test observations will be automatically classified as to have arisen from one of the known groups encountered in the training set, without the possibility of detecting previously unseen classes. To overcome this limitation, an adaptive semi-parametric Bayesian classifier is introduced for modeling the test units, where robust knowledge is extracted from the training set and incorporated within the priors’ model specification. A successful application of the proposed approach in a real-world problem is addressed.

Several contributions to the recent literature have shown that supervised learning is greatly enhanced when only the most relevant features are selected for building the discrimination rule. Unfortunately, outliers and wrongly labelled units may undermine the determination of relevant predictors, and almost no dedicated methodologies have been developed to face this issue. In the present paper, we introduce a new robust variable selection approach, that embeds a classifier within a greedy-forward procedure. An experiment on synthetic data is provided, to under- line the benefits of the proposed method in comparison with non-robust solutions.

Bissiri et al. (2016) present a general Bayesian approach where the like- lihood is replaced more generally by a loss function, which is used to derive the posterior distribution from the prior. In this paper, we exploit this idea for learning the structure of undirected graphical models over discrete variables.

Bayesian mixture models provide rich and flexible class tools which are particularly useful when there is unobserved heterogeneity in the data. When the number of subpopulations, called components, is assumed random, we allow the data to determine the complexity of the model. The latter property allows us to include a finite mixture model with a random number of components into the wider class of Bayesian nonparametric models. In this paper we consider multivariate discrete data, so that the class of mixtures is also referred to as latent class models. In particular, we let the number of latent classes to be random, and resort to Bayesian nonparametric techniques to devise a MCMC algorithm. The model is illustrated on an benchmark application dealing with role conflict

Radon is a noble gas that occurs in nature as a decay product of uranium. Radon is the principal contributor to natural background radiation and is considered to be one of the major leading causes of lung cancer. The main concern revolves around indoor environments where radon accumulates and reaches high concentrations. In this paper, a semiparametric random-effect M-quantile model is introduced to model radon concentration inside a building, and a way to estimate the model within the framework of robust maximum likelihood is presented. Using data collected in a monitoring survey carried out in the Lombardy Region (Italy) in 2003–2004, we investigate the impact of a number of factors, such as geological typologies of the soil and building characteristics, on indoor concentration. The proposed methodology permits the identification of building typologies prone to a high concentration of the pollutant. It is shown how these effects are largely not constant across the entire distribution of indoor radon concentration, making the suggested approach preferable to ordinary regression techniques since high concentrations are usually of concern. Furthermore, we demonstrate how our model provides a natural way of identifying those areas more prone to high concentration, displaying them by thematic maps. Understanding how buildings’ characteristics affect indoor concentration is fundamental both for preventing the gas from accumulating in new buildings and for mitigating those situations where the amount of radon detected inside a building is too high and has to be reduced.

A standard assumption when modelling linked sample data is that the stochastic properties of the linking process and process underpinning the population values of the response variable are independent of one another. This is often referred to as non-informative linkage. But what if linkage errors are informative? In this paper, we provide results from two simulation experiments that explore two potential informative linking scenarios. The first is where the choice of sample record to link is dependent on the response; and the second is where the probability of correct linkage is dependent on the response. We focus on the important and widely applicable problem of estimation of domain means given linked data, and provide empirical evidence that while standard domain estimation methods can be substantially biased in the presence of informative linkage errors, an alternative estimation method, based on a Gaussian approximation to a maximum likelihood estimator that allows for non-informative linkage error, performs well.

Health-related quality of life assessment is important in the clinical evaluation of patients with metastatic disease that may offer useful information in understanding the clinical effectiveness of a treatment. To assess if a set of explicative variables impacts on the health-related quality of life, regression models are routinely adopted. However, the interest of researchers may be focussed on modelling other parts (e.g. quantiles) of this conditional distribution. In this paper, we present an approach based on quantile and M-quantile regression to achieve this goal. We applied the methodologies to a prospective, randomized, multi-centre clinical trial. In order to take into account the hierarchical nature of the data we extended the M-quantile regression model to a three-level random effects specification and estimated it by maximum likelihood