Ricerc@Sapienza

Classification of multivariate time series (MTS) has been tackled with a large variety of methodologies and applied to a wide range of scenarios. Reservoir computing (RC) provides efficient tools to generate a vectorial, fixed-size representation of the MTS that can be further processed by standard classifiers. Despite their unrivaled training speed, MTS classifiers based on a standard RC architecture fail to achieve the same accuracy of fully trainable neural networks.

Gli algoritmi di apprendimento approfondito costituiscono una vasta classe di algoritmi machine learning basati sulla rappresentazione dei dati tramite molteplici livelli di astrazione. Essi sono stati applicati con successo in diverse aree

We present a model and a computational procedure for dealing with seasonality and regime changes in time series. In this work we are interested in time series which in addition to trend display seasonality in mean, in autocorrelation and in variance. These type of series appears in many areas, including hydrology, meteorology, economics and finance. The seasonality is accounted for by subset PAR modelling, for which each season follows a possibly different Autoregressive model.

Many white noise and goodness of fit tests are (asymptotically) written as quadratic forms in the ordinary autocorrelation estimates. The properties of such tests are studied by investigating the structure of the matrix of the quadratic form. We suggest to choose the matrix of the quadratic form in such a way that the power is maximized according to the information available about the alternative hypothesis. A simulation study sheds some light on the behavior of the test in finite samples.

We consider the problem of detecting a change in an arbitrary vector process by examining the evolution of calculated data subspaces. In our developments, both the data subspaces and the change identification criterion are novel and founded in the theory of L1-norm principal-component analysis (PCA). The outcome is highly accurate, rapid detection of change in streaming data that vastly outperforms conventional eigenvector subspace methods (L2-norm PCA).

The data provided by the National Institute for Statistics showed that, over the last ten years, the youth employment rate (15-34 years) has decreased by 10.2 percentage points. However, if we look at the segments, the young unemployment rate has considerably increased both for the young adults 15-24 years of age (from 20.4% in 2007 to 34.7% in 2017) and for the young adults 25-34 years (from 8.3% to 17%).

Our ability to predict the identity of future invasive alien species is largely based upon knowledge of prior invasion history. Emerging alien species-those never encountered as aliens before-therefore pose a significant challenge to biosecurity interventions worldwide. Understanding their temporal trends, origins, and the drivers of their spread is pivotal to improving prevention and risk assessment tools. Here, we use a database of 45,984 first records of 16,019 established alien species to investigate the temporal dynamics of occurrences of emerging alien species worldwide.

One of the measures that can reduce the negative effects of road accidents is the quick arrive of emergency vehicles to the accident area. This measure requires an effective location in space and on time of these vehicles. This location can be decided after an analysis of the available data in order to find the spatial and temporal characteristics of road accidents. The study presented in this paper uses time series accident data of the 15 districts of Rome Municipality, collected in four months in 2016.