1,721,066 research outputs found
Including covariates in a space-time point process with application to seismicity
Full text linkThe paper proposes a stochastic process that improves the assessment of
events in space and time, considering a contagion model (branching process) within
a regression-like framework to take covariates into account. The proposed approach
develops the Forward Likelihood for prediction (FLP) method for estimating the ETAS
model, including covariates in the model specification of the epidemic component. A
simulation study is carried out for analysing the misspecification model effect under
several scenarios. Also an application to the Italian seismic catalogue is reported,
together with the reference to the developed R packag
Statistical evaluation systems at 360°: techniques, technologies and new frontiers
No full textThis Special Issue has drawn inspiration from the Conference “Innovation and Society 2019—Statistical evaluation systems at 360°: techniques, technologies and new frontiers” (IES2019), which represents the 9th Scientific Meeting of the Statistics for the Evaluation and Quality of Services Group of the Italian Statistical Society—(SVQS), organized on 4 and 5 July 2019 at the “Università Europea di Roma” (Italy). This Conference was held with the aim of strengthening statistical research on evaluation topics, with a particular focus on economic and social issues, and highlighting how statistical thinking, design and analysis play a crucial role in social life, as well as how useful they are to society as a whole. To make the IES2019 conference as effective and constructive as possible, participation was open not only to scholars from different disciplines, but also to experts and decision-makers, who dealt with the relationship among evaluation, innovation and social issues.
The accomplishment of this issue aims at covering an extensive array of topics on evaluation systems of the public sector through statistical methods and models. Our goal was to provide a comprehensive technical, theoretical and practical description of a variety of statistical methodologies on evaluation issue.
The 11 articles of this Special Issue, selected after double-blind peer reviews, concern scientific studies applied in several fields, using many different statistical approaches and sharing the common aim stated by the Conference IES2019. From an empirical point of view, the articles deal with topics such as Education (Camminatiello et al., Cavicchia and Sarnacchiaro); Transport (D’Ambra et al., Iannario and Monti), Policy evaluation (De Iaco and Maggio, Pagani and Panarello), Economy and finance (Biffingandi and Zeli, Giacalone) and Social media (Mariani and Marletta). As regards methodological approaches, the statistical methods and models used are Regression models (Davino et al., De Iaco and Maggio, Iannario and Monti, Pagani and Panarello), Data analysis models (Camminatiello et al., Cavicchia and Sarnacchiaro, D’Ambra et al.), Techniques for building panels from archives, Missing data disambiguation, Assessing ranking dissimilarity and inter-group heterogeneity, Improving forecasting accuracy, Simulated data (Biffignandi and Zeli, Davino et al., Giacalone, Mariani and Marletta, Vanacore and Pellegrino). The heterogeneity of application and statistical approaches underline the broad spectrum of data analyses and the richness of the methodologies
Model selection for mixture hidden Markov models: an application to clickstream data
Full text linkIn a clickstream analysis setting, Mixture Hidden Markov Models (MHMMs) can be used to examine categorical sequences assuming they evolve according to a mixture of latent Markov processes, each related to a different subpopulation. These models involve identifying both the number of subpopulations and hidden states. This study proposes a model selection criterion based on an integrated completed likelihood approach that accounts for the two latent classes in the model.We implemented a Monte Carlo simulation study to compare selection criteria performance. In scenarios characterised by categorical short length sequences, our proposed measure outperforms the most commonly used model selection criteria in identifying components and states. The paper presents a case study on clickstream data collected from the website of a company operating in the hospitality industry and modelled by an MHMM selected by the proposed score
Demographic changes, research questions and data needs: issues about migrations
No full textAfter a synthetic introductory analysis on the trends of legal and resident population in Italy according to different sources, such as census, post-census estimations and Municipal Population Registers, the paper aims at focusing on the migratory component of population dynamics. In particular, it underlines the recent changes occurred to foreign immigration (for example, the importance of family reunifications and the growth of asylum seekers) to foreign-origin presence (from first to second generation, and from foreigners to new Italians), as well as to Italian emigration. This background framework will outline new information and research needs to which Italian public statistics system could or should handle considering the gained experience in other European countries (from cross-sectional to longitudinal surveys)
A penalized approach to covariate selection through quantile regression coefficient models
No full textThe coefficients of a quantile regression model are one-to-one functions of the order of the quantile. In standard quantile regression (QR), different quantiles are estimated one at a time. Another possibility is to model the coefficient functions parametrically, an approach that is referred to as quantile regression coefficients modeling (QRCM). Compared with standard QR, the QRCM approach facilitates estimation, inference and interpretation of the results, and generates more efficient estimators. We designed a penalized method that can address the selection of covariates in this particular modelling framework. Unlike standard penalized quantile regression estimators, in which model selection is quantile-specific, our approach permits using information on all quantiles simultaneously. We describe the estimator, provide simulation results and analyse the data that motivated the present article. The proposed approach is implemented in the qrcmNP package in R
Resilient students with migratory background
No full textNumerous research has shown that there is direct relationship among high educational outcomes, employment, income and integration. Immigrants’ children do show, on average, lower educational performances than natives’ ones in addition of being more socio-economically disadvantaged. Countries’ success in helping and integrating immigrants’ children needs to find new educational policies and strategies. Using PISA survey data, we aim at providing further elements of discussion looking at “resilient” students. Given the strong impact of the students’ social, cultural and economic family background, we aim at analysing how “disadvantaged” students can overcome their socio-economic and cultural obstacles and achieve good educational performances. Applying multilevel logit models, we analyse the main determinants to be resilient students considering both individual motivations and school resources
Stochastic Dominance for Generalized Parametric Families
No full textThe T-X family is a recent method for generating distributions by composing probability distributions and quantile functions. Such an approach makes it possible to obtain a large number of flexible families of parametric distributions,
new or already existing, most of which are typically used to model phenomena in different areas, such as economics and finance. We present a general method to derive sufficient conditions for the second-order stochastic dominance, within T-X families of distributions
Clustering Data Streams via Functional Data Analysis: a Comparison between Hierarchical Clustering and K-means Approaches
No full textRecently, the analysis of web data, has become essential in many research fields. For example, for a large number of companies, corporate strategies should
be based on the analysis of customer behaviour in surfing the world wide web. The main issues in analysing web traffic and web data are that they often flow continu
ously from a source and are potentially unbounded in size, and these circumstances inhibit to store the whole dataset. In this paper, we propose an alternative clustering
functional data stream method to implement existing techniques, and we address phenomena in which web data are expressed by a curve or a function. In particu-
lar, we deal with a specific type of web data, i.e. trends of google queries. Specifically, focusing on top football players data, we compare the functional k-means
approach to the functional Hierarchical Clustering for detecting specific pattern of search trends over time
Is Structural Equation Modelling Able to Predict Well-being? / È possibile stimare il livello di benessere per mezzo di modelli ad equazioni strutturali?
No full textThe well-being (WB) measurement is an important and challenging task. Quality of life is a multifaceted topic, thus its measure cannot rely anymore on one or few indicators only. Data from large-scale survey projects, such as the European Social Survey (ESS), are a solid basis for testing new methods aimed at measuring the phenomenon. We apply Structural Equation Modelling (SEM) to ESS wave 8 data. Our research aims at evaluating if SEM is a reliable method for estimating the WB and the relative importance of its dimensions in some European countries
Benefits of the Erasmus mobility experience: a discrete latent variable analysis
No full textInternationalization of higher education has become a priority in the European education policy. For this reason, research in this area is expanding with the aim of understanding motivations and potential benefits of international mobility. In such a context, an online survey addressed to about 1,600 students with Erasmus mobility experiences was conducted by the University of Bergamo (IT). Two latent traits, that is, the impact of the Erasmus experience on the student’s abilities and the student’s satisfaction for this experience, are analyzed through a two-dimensional latent class Item Response Theory model under a concomitant variable approach. The twofold issue concerning the choice of the optimal number of latent classes as well as the selection of significant covariates is specially addressed
- …
