This paper is devoted to the R package fda.usc which includes some utilities for functional data analysis. This package carries out exploratory and descriptive analysis of functional data analyzing its most important features such as depth measurements or functional outliers detection, among others. The R package fda.usc also includes functions to compute functional regression models, with a scalar response and a functional explanatory data via non-parametric functional regression, basis representation or functional principal components analysis. There are natural extensions such as functional linear models and semi-functional partial linear models, which allow non-functional covariates and factors and make predictions. The functions of this package complement and incorporate the two main references of functional data analysis: The R package fda and the functions implemented by Ferraty and Vieu (2006).

The Maximum Depth classifier was the first attempt to use data depths instead of multivariate raw data in classification problems. Recently, the DD-classifier has fixed some serious limitations of this classifier but some issues still remain. This paper is devoted to extending the DD-classifier in the following ways: first, to be able to handle more than two groups; second, to apply regular classification methods (such as kNN, linear or quadratic classifiers, recursive partitioning,. . . ) to DD-plots, which, in particular, allows to obtain useful insights through the diagnostics of these methods; and third, to integrate various sources of information (data depths, multivariate functional data,. . . ) in the classification procedure in an unified way. An enhanced revision of several functional data depths is also proposed. A simulation study and applications to some real datasets are also provided.

This paper considers the problem of variable selection in regression models in the case of functional variables that may be mixed with other type of variables (scalar, multivariate, directional, etc.). Our proposal begins with a simple null model and sequentially selects a new variable to be incorporated into the model based on the use of distance correlation proposed by Székely et al. [2007]. For the sake of simplicity, this paper only uses additive models. However, the proposed algorithm may assess the type of contribution (linear, non linear, ...) of each variable. The algorithm has shown quite promising results when applied to simulations and real data sets.

Influenza surveillance is critical to monitoring the situation during epidemic seasons and predictive mathematic models may aid the early detection of epidemic patterns. The objective of this study was to design a real-time spatial predictive model of ILI (Influenza Like Illness) incidence rate in Catalonia using one- and two-week forecasts. The available data sources used to select explanatory variables to include in the model were the statutory reporting disease system and the sentinel surveillance system in Catalonia for influenza incidence rates, the official climate service in Catalonia for meteorological data, laboratory data and Google Flu Trend. Time series for every explanatory variable with data from the last 4 seasons (from 2010–2011 to 2013–2014) was created. A pilot test was conducted during the 2014–2015 season to select the explanatory variables to be included in the model and the type of model to be applied. During the 2015–2016 season a real-time model was applied weekly, obtaining the intensity level and predicted incidence rates with 95% confidence levels one and two weeks away for each health region. At the end of the season, the confidence interval success rate (CISR) and intensity level success rate (ILSR) were analysed. For the 2015–2016 season a CISR of 85.3% at one week and 87.1% at two weeks and an ILSR of 82.9% and 82% were observed, respectively. The model described is a useful tool although it is hard to evaluate due to uncertainty. The accuracy of prediction at one and two weeks was above 80% globally, but was lower during the peak epidemic period. In order to improve the predictive power, new explanatory variables should be included.

This paper proposes a novel approach that uses meteorological information to predict the incidence of influenza in Galicia (Spain). It extends the Generalized Least Squares (GLS) methods in the multivariate framework to functional regression models with dependent errors. These kinds of models are useful when the recent history of the incidence of influenza are readily unavailable (for instance, by delays on the communication with health informants) and the prediction must be constructed by correcting the temporal dependence of the residuals and using more accessible variables. A simulation study shows that the GLS estimators render better estimations of the parameters associated with the regression model than they do with the classical models. They obtain extremely good results from the predictive point of view and are competitive with the classical time series approach for the incidence of influenza. An iterative version of the GLS estimator (called iGLS) was also proposed that can help to model complicated dependence structures. For constructing the model, the distance correlation measure was employed to select relevant information to predict influenza rate mixing multivariate and functional variables. These kinds of models are extremely useful to health managers in allocating resources in advance to manage influenza epidemics.