206 research outputs found
Local indicators of spatio-temporal association on linear networks
No full textIn this work, we extend the Local Indicators of Spatio-Temporal Association (LISTA) functions (Siino et al. 2018) to the non-Euclidean space of linear networks. We introduce the local version of some inhomogeneous second-order statistics for spatio-temporal point processes on linear networks (Morandi and Mateu, 2019), namely the K-function and the pair correlation function.
Following the work of Adelfio et al. (2019) for the Euclidean case, we employ the proposed LISTA functions to assess the goodness-of-fit of different spatio-temporal models fitted to point patterns occurring on linear networks. Indeed, the peculiar lack of homogeneity in a network discourages the usage of traditional spatial and spatio-temporal methods based on stationary processes. Therefore, the weighted second-order statistics are appropriate diagnostic tools since they directly apply to data without assuming homogeneity.
We provide simulation studies, by generating both inhomogeneous and self-exiting spatio-temporal point processes on networks, and by carrying out diagnostics on different fitted intensities. By comparing the values of the LISTA functions and their theoretical values, we show that the LISTA can correctly identify the true intensity when this is constrained on a network
stopp: An R Package for Spatio-Temporal Point Pattern Analysis
No full textstopp is a novel R package specifically designed for the analysis of spatio-temporal point patterns which might have occurred in a subset of the Euclidean space or on some specific linear network, such as roads of a city. It represents the first package providing a comprehensive modelling framework for spatio-temporal Poisson point processes. While many specialized models exist in the scientific literature for analyzing complex spatio-temporal point patterns, we address the lack of general software for comparing simpler alternative models and their goodness of fit. The package’s main functionalities include modelling and diagnostics, together with exploratory analysis tools and the simulation of point processes. A particular focus is given to local first-order and second-order characteristics. The package aggregates existing methods within one coherent framework, including those we proposed in recent papers, and it aims to welcome many further proposals and extensions from the R community
Hints of latent drivers investigating university student performance
No full textJob market, nowadays, asks for higher and higher skills and competences. Therefore, also the measurement and assessment of the university students performance are crucial issues for policy makers. Although the scientific literature provides several papers investigating the main determinants of university student performance, often results are very different, and they seem to hold just in very specific contexts. This paper aims to contribute to the international literature, focusing on the role of student specific characteristics, supporting the idea that unobservable variables (such as motivation, aptitudes or abilities) should be more investigated
Minimum contrast for the first-order intensity estimation of spatial and spatio-temporal point processes
No full textIn this paper, we harness a result in point process theory, specifically the expectation of the weighted K-function, where the weighting is done by the true first-order intensity function. This theoretical result can be employed as an estimation method to derive parameter estimates for a particular model assumed for the data. The underlying motivation is to avoid the difficulties associated with dealing with complex likelihoods in point process models and their maximization. The exploited result makes our method theoretically applicable to any model specification. In this paper, we restrict our study to Poisson models, whose likelihood represents the base for many more complex point process models. In this context, our proposed method can estimate the vector of local parameters that correspond to the points within the analyzed point pattern without introducing any additional complexity compared to the global estimation. We illustrate the method through simulation studies for both purely spatial and spatio-temporal
point processes, and show complex scenarios based on the Poisson model through
the analysis of two real datasets concerning environmental problems
Advanced spatio-temporal point processes for the Sicily seismicity analysis
No full textDue to the complexity of the generator process of seismic events, we
study under several aspects the interaction structure between earthquake events using
recently developed spatio-temporal statistical techniques and models. Using
these advanced statistical tools, we aim to characterise the global and local scale
cluster behaviour of the Easter Sicily seismicity considering the catalogue data since
2006, when the Italian National Seismic Network was upgraded and earthquake location
was sensibly improved. Firstly, we characterise the global complex spatiotemporal
interaction structure with the space-time ETAS model where background
seismicity is estimated non-parametrically, while triggered seismicity is estimated
by MLE. After identifying seismic sequences by a clustering technique, we characterise
their spatial and spatio-temporal interaction structures using other advanced
point process models. For the characterisation of the spatial interactions, a version
of hybrid of Gibbs point process models is proposed as method to describe the
multiscale interaction structure of several seismic sequences accounting for both
the attractive and repulsive nature of data. Furthermore, we consider log-Gaussian
Cox processes (LGCP), that are relatively tractable class of empirical models for
describing spatio-temporal correlated phenomena. Several parametric formulation
of spatio-temporal LGCP are estimated, by the minimum contrast procedure, assuming
both separable and non-separable parametric specification of the correlation
function of the underlying Gaussian Random Field
Minimum contrast for estimating point processes intensity
No full textA result in point process theory, based on the expectation of the weighted K-function, is exploited by the true first-order intensity function. This
theoretical result can be an estimation method for obtaining the parameter estimates of a specific model assumed for the data. The motivation is
to avoid dealing with the complex likelihoods of some complex point process models and their maximization. This can be more evident when
considering the local second-order characteristics since the proposed method can estimate the vector of the local parameters corresponding to the
points of the analysed point pattern. We illustrate the method through simulation studies for purely spatial and spatio-temporal point processes
Including covariates in a space-time point process with application to seismicity
No full textThe 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
Selecting the Kth nearest-neighbour for clutter removal in spatial point processes through segmented regression models
No full textWe consider the problem of feature detection, in the presence of clutter in spatial point processes. A previous study addresses the issue of the selection of the best nearest neighbour for clutter removal. We outline a simple workflow to automatically estimate the number of nearest neighbours by means of segmented regression models applied to an entropy measure of cluster separation. The method is suitable for a feature with clutter as two superimposed Poisson processes on any twodimensional space, including linear networks. We present simulations to illustrate the method and an application to the problem of seismic fault detection
Minimum contrast for point processes' first-order intensity estimation
No full textIn this paper, we exploit some theoretical results, from which we know the expected value of the K-function weighted by the true first-order intensity function of a point pattern. This theoretical result can serve as an estimation method for obtaining the parameter estimates of a specific model, assumed for the data. The only requirement is the knowledge of the first-order intensity function expression, completely avoiding writing the likelihood, which is often complex to deal with in point process models. We illustrate the method through simulation studies for spatio-temporal point processes
- …
