Antonella Plaia - Academia.edu (original) (raw)

Papers by Antonella Plaia

Research paper thumbnail of A weighted distance-based approach with boosted decision trees for label ranking

Expert Systems with Applications

Research paper thumbnail of A family of distances for preference–approvals

Annals of Operations Research

A preference–approval on a set of alternatives consists of a weak order on that set and, addition... more A preference–approval on a set of alternatives consists of a weak order on that set and, additionally, a cut-off line that separates acceptable and unacceptable alternatives. In this paper, we propose a new method for defining the distance between preference–approvals taking into account jointly the disagreements in preferences and approvals for each pair of alternatives. The proposed distance is compared to the existing distance functions to deal with clustering problems. Specifically, we prove that our metric improves the estimated clusters in terms of both stability and accuracy.

Research paper thumbnail of Comparison of Tests for Heteroscedasticity in Linear Models

Proceedings of the 2022 AERA Annual Meeting

Research paper thumbnail of Classification trees for preference data: a distance-based approach

Research paper thumbnail of Citation

This article was submitted to Crop

Research paper thumbnail of A Projection Pursuit Algorithm for Preference Data

It is our pleasure to welcome the guests, participants and contributors to the International Conf... more It is our pleasure to welcome the guests, participants and contributors to the International Conference (SMTDA 2018) on Stochastic Modeling Techniques and Data Analysis and (DEMOGRAPHICS2018) Demographic Analysis and Research Workshop. The main goal of the conference is to promote new methods and techniques for analyzing data, in fields like stochastic modeling, optimization techniques, statistical methods and inference, data mining and knowledge systems, computing-aided decision supports, neural networks, chaotic data analysis, demography and life table data analysis. SMTDA Conference and DEMOGRAPHICS Workshop aim at bringing together people from both stochastic, data analysis and demography areas. Special attention is given to applications or to new theoretical results having potential of solving real life problems. SMTDA 2018 and DEMOGRAPHICS 2018 focus in expanding the development of the theories, the methods and the empirical data and computer techniques, and the best theoretical achievements of the Stochastic Modeling Techniques and Data Analysis field, bringing together various working groups for exchanging views and reporting research findings.

Research paper thumbnail of Aggregate air pollution indices: a new proposal

A new aggregate Air Quality Index (I2) to represent the global air pollution situation for a give... more A new aggregate Air Quality Index (I2) to represent the global air pollution situation for a given city/region is proposed. Accounting for simultaneous exposure to common pollutants and their effects on human health, this index overcomes existing AQIs. Its goodness and utility is shown by a simulation plan and by an application to a real dataset on main pollutants

Research paper thumbnail of Missing Data in Space-time: Long Gaps Imputation Based On Functional Data Analysis

The joint TIES-GRASPA 2017 Conference on Climate and Environment is held at the University of Ber... more The joint TIES-GRASPA 2017 Conference on Climate and Environment is held at the University of Bergamo, Italy from 24th to 26th July and it is a satellite meeting of the 61st World Statistics Congress-ISI2017 in Marrakesh. The International Environmetrics Society (TIES) is a non-profit organization aimed to foster the development and use of statistical and other quantitative methods in the environmental sciences, environmental engineering and environmental monitoring and protection. The Society promotes the participation of statisticians, mathematicians, scientists and engineers in the solution of environmental problems and emphasizes the need for collaboration and for clear communication between individuals from different disciplines and between researchers and practitioners. The Italian environmetricians group named GRASPA is active since 1995 and it has become since May 2013 a standing group of the Italian Statistical Society (SIS) for Environmental Statistics, sustainability and territorial safety. GRASPA-SIS promotes statistical and interdisciplinary research in the field of environmental quality, safety and sustainability including air and water quality, epidemiology, earth science and ecology. The conference is intended to be a bridge for a future TIES conference system after ISI incorporation. On the one side, it renews a fruitful tradition of joint conferences with other scientific societies and associations. On the other side, it explores new ways to collaborate with ISI. From the scientific point of view, the TIES-GRASPA conference on Climate and Environment is focused on a hot topic and has the potential to attract new statisticians and other scientists.

Research paper thumbnail of A new OLS-based procedure for clusterwise linear regression

Data heterogeneity, within a (linear) regression framework, often suggests the use of a Clusterwi... more Data heterogeneity, within a (linear) regression framework, often suggests the use of a Clusterwise Linear Regression (CLR) procedure, which implies, among other things, the estimate of the appropriate number of clusters as well as the cluster membership of each unit. The approaches to the estimation of a CLR model are essentially based on the Ordinary Least Square (OLS) criterion or the likelihood criterion. In this paper, in a context of OLS approach, we propose an estimation of the model making use of an algorithm based on a threshold criterion for the determination coefficient of each cluster, to identify the appropriate number of clusters, and of a modified Spath's algorithm, to estimate the cluster membership of each sample unit. A simulation design and an application to a real data-set show that the procedure outperforms other algorithms commonly used in the literature

Research paper thumbnail of An aggregate air quality index considering interactions among pollutants

Several countries provide an Air Quality Index (AQI) to communicate air pollution, but there is n... more Several countries provide an Air Quality Index (AQI) to communicate air pollution, but there is not a unique and nternationally accepted methodology for constructing it. The most of the proposed indices are based on the USA AQI by EPA and are defined by the value of the pollutant with the highest concentration. For each pollutant, a sub-index is computed by linear interpolation according to the grid in a table, but the breakpoints of such a table may differ from one country to another, as well as the descriptors of each category, the air quality standards, the functions chosen as daily synthesis to aggregate hourly values at each site for each pollutant, and so on. Anyway the main drawback is that such indices do not take into account the combined effects of all the considered pollutants, giving little emphasis to effects occurring over long time periods, such as chronic health effects, damages on vegetation, animals, monuments. With the purpose to account for multiple pollutant exposure, some attempts have already been made in literature. In this paper we propose an aggregate AQI which tries to overcome the problems listed above. The proposed index can be easily computed and interpreted, so allowing comparisons among different situations, both in time and space. Moreover, the association of an appropriate measure of dispersion, that can also be considered as a measure of concentration, adds very important information

Research paper thumbnail of Weighted and unweighted distances based decision tree for ranking data

Preference data represent a particular type of ranking data (widely used in sports, web search, s... more Preference data represent a particular type of ranking data (widely used in sports, web search, social sciences), where a group of people gives their preferences over a set of alternatives. Within this framework, distance-based decision trees represent a non-parametric tool for identifying the profiles of subjects giving a similar ranking. This paper aims at detecting, in the framework of (complete and incomplete) ranking data, the impact of the differently structured weighted distances for building decision trees. The traditional metrics between rankings don’t take into account the importance of swapping elements similar among them (element weights) or elements belonging to the top (or to the bottom) of an ordering (position weights). By means of simulations, using weighted distances to build decision trees, we will compute the impact of different weighting structures both on splitting and on consensus ranking. The distances that will be used satisfy Kemenys axioms and, accordingly...

Research paper thumbnail of Item weighted Kemeny distance for preference data

Preference data represent a particular type of ranking data where a group of people gives their p... more Preference data represent a particular type of ranking data where a group of people gives their preferences over a set of alternatives. The traditional metrics between rankings don't take into account that the importance of elements can be not uniform. In this paper the item weighted Kemeny distance is introduced and its properties demonstrated.

Research paper thumbnail of Element weighted Kemeny distance for ranking data

Electronic Journal of Applied Statistical Analysis, 2021

Preference data are a particular type of ranking data that arise when n individuals express their... more Preference data are a particular type of ranking data that arise when n individuals express their preferences over a finite set of items. Within this framework, the main issue concerns the aggregation of the preferences to identify a compromise or a “consensus”, defined as the closest ranking (i.e. with the minimum distance or maximum correlation) to the whole set of preferences. Many approaches have been proposed, but they are not sensitive to the importance of items: i.e. changing the rank of a highly-relevant element should result in a higher penalty than changing the rank of a negligible one. The goal of this paper is to investigate the consensus between rankings taking into account the importance of items (element weights). For this purpose, we present: i) an element weighted rank correlation coefficient tau_ew as an extension of the Emond and Mason’s tau, and ii) an element weighted rank distance d_ew as an extension of the Kemeny distance d. The one-to-one correspondence ...

Research paper thumbnail of Principal components for multivariate spatiotemporal functional data

Research paper thumbnail of GAMs and functional kriging for air quality data

In environmental sciences data having spatio-temporal structure are often observed. They can be c... more In environmental sciences data having spatio-temporal structure are often observed. They can be considered as discrete observations from curves along time and/or space and treated as functional. Generalized Additive Models (GAMs)represent a useful tool for modelling, for example, pollutant concentrations and describing their spatial and/or temporal trends.Very often the prediction of a curve at an unmonitored site is necessary. At this aim we extend kriging for functional data to a multivariate context. Actually, even if we are interested only in predicting a single pollutant, for example PM10, exploiting its correlation with the other pollutants can improve the estimation. Cross validation is used to test the performance of the proposed procedure.

Research paper thumbnail of Comparing spatial and spatio-temporal FPCA to impute large continuous gaps in space

Multivariate spatio-temporal data analysis methods usually assume fairly complete data, while a n... more Multivariate spatio-temporal data analysis methods usually assume fairly complete data, while a number of gaps often occur along time or in space. In air quality data long gaps may be due to instrument malfunctions; moreover, not all the pollutants of interest are measured in all the monitoring stations of a network. In literature, many statistical methods have been proposed for imputing short sequences of missing values, but most of them are not valid when the fraction of missing values is high. Furthermore, the limitation of the methods commonly used consists in exploiting temporal only, or spatial only, correlation of the data. The objective of this paper is to provide an approach based on spatio-temporal functional principal component analysis (FPCA), exploiting simultaneously the spatial and temporal correlations for multivariate data, in order to provide an accurate imputation of missing values. At this aim, the methodology proposed in a previous proposal is applied, in order ...

Research paper thumbnail of Ensemble Methods for Ranking Data

The last years have seen a remarkable flowering of works about the use of decision trees for rank... more The last years have seen a remarkable flowering of works about the use of decision trees for ranking data. As a matter of fact, decision trees are useful and intuitive, but they are very unstable: small perturbations bring big changes. This is the reason why it could be necessary to use more stable procedures, as ensemble methods, in order to find which predictors are able to explain the preference structure. In this work ensemble methods as BAGGING and Random Forest are proposed, from both a theoretical and computational point of view, for deriving classification trees when ranking data are observed. The advantages of these procedures are shown through an example on the SUSHI data set

Research paper thumbnail of Functional principal component analysis of quantile curves

Literature on functional data analysis is mainly focused on estimation of individuals curves and... more Literature on functional data analysis is mainly focused on estimation of individuals curves and characterization of average dynamics. The idea underlying this proposal is to focus attention on other particular features of the distribution of the observed data, moving from mean functions towards functional quantiles. The motivating examples are functional data sets that are collections of high frequency data recorded along time. As quantiles provide information on various aspects of a time series, we propose a modelling framework for the joint estimation of functional quantiles, varying along time, and functional principal components, summarizing some common dynamics shared by the functional quantiles.

Research paper thumbnail of Can the students' career performance be helpful in predicting an increase in universities income?

The students\u2019 academic failure and the delay in obtaining their final degree are a significa... more The students\u2019 academic failure and the delay in obtaining their final degree are a significant issue for the Italian universities and their shareholders. Based on indicators proposed by the Italian Ministry of University, the Italian universities are awarded a financial incentive if they reduce the students\u2019 attrition and failure. In this paper we analyze the students\u2019 careers performance using: 1) aggregate data; 2) individual data. The first compares the performances of the Italian universities using the measures and the indicators proposed by the Ministry. The second analyzes the students\u2019 careers through an indicator based on credit earned by each student in seven academic years. The primary goal of this paper is to highlight elements that can be used by the policy makers to improve the careers of the university students

Research paper thumbnail of L Ong Gaps in Multivariate Spatio-Temporal Data : An Approach Based on F Unctional D Ata a Nalysis

The main aim of this paper is to perform Functional Principal Component Analysis (FPCA) taking in... more The main aim of this paper is to perform Functional Principal Component Analysis (FPCA) taking into account spatio-temporal correlation structures, in order to fill in missing values in spatio-temporal multivariate data set.A spatial and a spatio-temporal variant of the classical temporal FPCA is considered; in other words, FPCA is carried out after modeling data with respect to more than one dimension: space (long, lat) or space+time. Moreover, multidimensional FPCA is extended to multivariate context (more than one variable). Information on spatial or spatiotemporal structures are efficiently extracted by applying Generalized Additive Models (GAMs). Both simulation studies and some performance indicators are used to validate the proposed procedure, showing that, especially in presence of long gaps, spatio-temporal FPCA provides a better reconstruction than spatial FPCA.

Research paper thumbnail of A weighted distance-based approach with boosted decision trees for label ranking

Expert Systems with Applications

Research paper thumbnail of A family of distances for preference–approvals

Annals of Operations Research

A preference–approval on a set of alternatives consists of a weak order on that set and, addition... more A preference–approval on a set of alternatives consists of a weak order on that set and, additionally, a cut-off line that separates acceptable and unacceptable alternatives. In this paper, we propose a new method for defining the distance between preference–approvals taking into account jointly the disagreements in preferences and approvals for each pair of alternatives. The proposed distance is compared to the existing distance functions to deal with clustering problems. Specifically, we prove that our metric improves the estimated clusters in terms of both stability and accuracy.

Research paper thumbnail of Comparison of Tests for Heteroscedasticity in Linear Models

Proceedings of the 2022 AERA Annual Meeting

Research paper thumbnail of Classification trees for preference data: a distance-based approach

Research paper thumbnail of Citation

This article was submitted to Crop

Research paper thumbnail of A Projection Pursuit Algorithm for Preference Data

It is our pleasure to welcome the guests, participants and contributors to the International Conf... more It is our pleasure to welcome the guests, participants and contributors to the International Conference (SMTDA 2018) on Stochastic Modeling Techniques and Data Analysis and (DEMOGRAPHICS2018) Demographic Analysis and Research Workshop. The main goal of the conference is to promote new methods and techniques for analyzing data, in fields like stochastic modeling, optimization techniques, statistical methods and inference, data mining and knowledge systems, computing-aided decision supports, neural networks, chaotic data analysis, demography and life table data analysis. SMTDA Conference and DEMOGRAPHICS Workshop aim at bringing together people from both stochastic, data analysis and demography areas. Special attention is given to applications or to new theoretical results having potential of solving real life problems. SMTDA 2018 and DEMOGRAPHICS 2018 focus in expanding the development of the theories, the methods and the empirical data and computer techniques, and the best theoretical achievements of the Stochastic Modeling Techniques and Data Analysis field, bringing together various working groups for exchanging views and reporting research findings.

Research paper thumbnail of Aggregate air pollution indices: a new proposal

A new aggregate Air Quality Index (I2) to represent the global air pollution situation for a give... more A new aggregate Air Quality Index (I2) to represent the global air pollution situation for a given city/region is proposed. Accounting for simultaneous exposure to common pollutants and their effects on human health, this index overcomes existing AQIs. Its goodness and utility is shown by a simulation plan and by an application to a real dataset on main pollutants

Research paper thumbnail of Missing Data in Space-time: Long Gaps Imputation Based On Functional Data Analysis

The joint TIES-GRASPA 2017 Conference on Climate and Environment is held at the University of Ber... more The joint TIES-GRASPA 2017 Conference on Climate and Environment is held at the University of Bergamo, Italy from 24th to 26th July and it is a satellite meeting of the 61st World Statistics Congress-ISI2017 in Marrakesh. The International Environmetrics Society (TIES) is a non-profit organization aimed to foster the development and use of statistical and other quantitative methods in the environmental sciences, environmental engineering and environmental monitoring and protection. The Society promotes the participation of statisticians, mathematicians, scientists and engineers in the solution of environmental problems and emphasizes the need for collaboration and for clear communication between individuals from different disciplines and between researchers and practitioners. The Italian environmetricians group named GRASPA is active since 1995 and it has become since May 2013 a standing group of the Italian Statistical Society (SIS) for Environmental Statistics, sustainability and territorial safety. GRASPA-SIS promotes statistical and interdisciplinary research in the field of environmental quality, safety and sustainability including air and water quality, epidemiology, earth science and ecology. The conference is intended to be a bridge for a future TIES conference system after ISI incorporation. On the one side, it renews a fruitful tradition of joint conferences with other scientific societies and associations. On the other side, it explores new ways to collaborate with ISI. From the scientific point of view, the TIES-GRASPA conference on Climate and Environment is focused on a hot topic and has the potential to attract new statisticians and other scientists.

Research paper thumbnail of A new OLS-based procedure for clusterwise linear regression

Data heterogeneity, within a (linear) regression framework, often suggests the use of a Clusterwi... more Data heterogeneity, within a (linear) regression framework, often suggests the use of a Clusterwise Linear Regression (CLR) procedure, which implies, among other things, the estimate of the appropriate number of clusters as well as the cluster membership of each unit. The approaches to the estimation of a CLR model are essentially based on the Ordinary Least Square (OLS) criterion or the likelihood criterion. In this paper, in a context of OLS approach, we propose an estimation of the model making use of an algorithm based on a threshold criterion for the determination coefficient of each cluster, to identify the appropriate number of clusters, and of a modified Spath's algorithm, to estimate the cluster membership of each sample unit. A simulation design and an application to a real data-set show that the procedure outperforms other algorithms commonly used in the literature

Research paper thumbnail of An aggregate air quality index considering interactions among pollutants

Several countries provide an Air Quality Index (AQI) to communicate air pollution, but there is n... more Several countries provide an Air Quality Index (AQI) to communicate air pollution, but there is not a unique and nternationally accepted methodology for constructing it. The most of the proposed indices are based on the USA AQI by EPA and are defined by the value of the pollutant with the highest concentration. For each pollutant, a sub-index is computed by linear interpolation according to the grid in a table, but the breakpoints of such a table may differ from one country to another, as well as the descriptors of each category, the air quality standards, the functions chosen as daily synthesis to aggregate hourly values at each site for each pollutant, and so on. Anyway the main drawback is that such indices do not take into account the combined effects of all the considered pollutants, giving little emphasis to effects occurring over long time periods, such as chronic health effects, damages on vegetation, animals, monuments. With the purpose to account for multiple pollutant exposure, some attempts have already been made in literature. In this paper we propose an aggregate AQI which tries to overcome the problems listed above. The proposed index can be easily computed and interpreted, so allowing comparisons among different situations, both in time and space. Moreover, the association of an appropriate measure of dispersion, that can also be considered as a measure of concentration, adds very important information

Research paper thumbnail of Weighted and unweighted distances based decision tree for ranking data

Preference data represent a particular type of ranking data (widely used in sports, web search, s... more Preference data represent a particular type of ranking data (widely used in sports, web search, social sciences), where a group of people gives their preferences over a set of alternatives. Within this framework, distance-based decision trees represent a non-parametric tool for identifying the profiles of subjects giving a similar ranking. This paper aims at detecting, in the framework of (complete and incomplete) ranking data, the impact of the differently structured weighted distances for building decision trees. The traditional metrics between rankings don’t take into account the importance of swapping elements similar among them (element weights) or elements belonging to the top (or to the bottom) of an ordering (position weights). By means of simulations, using weighted distances to build decision trees, we will compute the impact of different weighting structures both on splitting and on consensus ranking. The distances that will be used satisfy Kemenys axioms and, accordingly...

Research paper thumbnail of Item weighted Kemeny distance for preference data

Preference data represent a particular type of ranking data where a group of people gives their p... more Preference data represent a particular type of ranking data where a group of people gives their preferences over a set of alternatives. The traditional metrics between rankings don't take into account that the importance of elements can be not uniform. In this paper the item weighted Kemeny distance is introduced and its properties demonstrated.

Research paper thumbnail of Element weighted Kemeny distance for ranking data

Electronic Journal of Applied Statistical Analysis, 2021

Preference data are a particular type of ranking data that arise when n individuals express their... more Preference data are a particular type of ranking data that arise when n individuals express their preferences over a finite set of items. Within this framework, the main issue concerns the aggregation of the preferences to identify a compromise or a “consensus”, defined as the closest ranking (i.e. with the minimum distance or maximum correlation) to the whole set of preferences. Many approaches have been proposed, but they are not sensitive to the importance of items: i.e. changing the rank of a highly-relevant element should result in a higher penalty than changing the rank of a negligible one. The goal of this paper is to investigate the consensus between rankings taking into account the importance of items (element weights). For this purpose, we present: i) an element weighted rank correlation coefficient tau_ew as an extension of the Emond and Mason’s tau, and ii) an element weighted rank distance d_ew as an extension of the Kemeny distance d. The one-to-one correspondence ...

Research paper thumbnail of Principal components for multivariate spatiotemporal functional data

Research paper thumbnail of GAMs and functional kriging for air quality data

In environmental sciences data having spatio-temporal structure are often observed. They can be c... more In environmental sciences data having spatio-temporal structure are often observed. They can be considered as discrete observations from curves along time and/or space and treated as functional. Generalized Additive Models (GAMs)represent a useful tool for modelling, for example, pollutant concentrations and describing their spatial and/or temporal trends.Very often the prediction of a curve at an unmonitored site is necessary. At this aim we extend kriging for functional data to a multivariate context. Actually, even if we are interested only in predicting a single pollutant, for example PM10, exploiting its correlation with the other pollutants can improve the estimation. Cross validation is used to test the performance of the proposed procedure.

Research paper thumbnail of Comparing spatial and spatio-temporal FPCA to impute large continuous gaps in space

Multivariate spatio-temporal data analysis methods usually assume fairly complete data, while a n... more Multivariate spatio-temporal data analysis methods usually assume fairly complete data, while a number of gaps often occur along time or in space. In air quality data long gaps may be due to instrument malfunctions; moreover, not all the pollutants of interest are measured in all the monitoring stations of a network. In literature, many statistical methods have been proposed for imputing short sequences of missing values, but most of them are not valid when the fraction of missing values is high. Furthermore, the limitation of the methods commonly used consists in exploiting temporal only, or spatial only, correlation of the data. The objective of this paper is to provide an approach based on spatio-temporal functional principal component analysis (FPCA), exploiting simultaneously the spatial and temporal correlations for multivariate data, in order to provide an accurate imputation of missing values. At this aim, the methodology proposed in a previous proposal is applied, in order ...

Research paper thumbnail of Ensemble Methods for Ranking Data

The last years have seen a remarkable flowering of works about the use of decision trees for rank... more The last years have seen a remarkable flowering of works about the use of decision trees for ranking data. As a matter of fact, decision trees are useful and intuitive, but they are very unstable: small perturbations bring big changes. This is the reason why it could be necessary to use more stable procedures, as ensemble methods, in order to find which predictors are able to explain the preference structure. In this work ensemble methods as BAGGING and Random Forest are proposed, from both a theoretical and computational point of view, for deriving classification trees when ranking data are observed. The advantages of these procedures are shown through an example on the SUSHI data set

Research paper thumbnail of Functional principal component analysis of quantile curves

Literature on functional data analysis is mainly focused on estimation of individuals curves and... more Literature on functional data analysis is mainly focused on estimation of individuals curves and characterization of average dynamics. The idea underlying this proposal is to focus attention on other particular features of the distribution of the observed data, moving from mean functions towards functional quantiles. The motivating examples are functional data sets that are collections of high frequency data recorded along time. As quantiles provide information on various aspects of a time series, we propose a modelling framework for the joint estimation of functional quantiles, varying along time, and functional principal components, summarizing some common dynamics shared by the functional quantiles.

Research paper thumbnail of Can the students' career performance be helpful in predicting an increase in universities income?

The students\u2019 academic failure and the delay in obtaining their final degree are a significa... more The students\u2019 academic failure and the delay in obtaining their final degree are a significant issue for the Italian universities and their shareholders. Based on indicators proposed by the Italian Ministry of University, the Italian universities are awarded a financial incentive if they reduce the students\u2019 attrition and failure. In this paper we analyze the students\u2019 careers performance using: 1) aggregate data; 2) individual data. The first compares the performances of the Italian universities using the measures and the indicators proposed by the Ministry. The second analyzes the students\u2019 careers through an indicator based on credit earned by each student in seven academic years. The primary goal of this paper is to highlight elements that can be used by the policy makers to improve the careers of the university students

Research paper thumbnail of L Ong Gaps in Multivariate Spatio-Temporal Data : An Approach Based on F Unctional D Ata a Nalysis

The main aim of this paper is to perform Functional Principal Component Analysis (FPCA) taking in... more The main aim of this paper is to perform Functional Principal Component Analysis (FPCA) taking into account spatio-temporal correlation structures, in order to fill in missing values in spatio-temporal multivariate data set.A spatial and a spatio-temporal variant of the classical temporal FPCA is considered; in other words, FPCA is carried out after modeling data with respect to more than one dimension: space (long, lat) or space+time. Moreover, multidimensional FPCA is extended to multivariate context (more than one variable). Information on spatial or spatiotemporal structures are efficiently extracted by applying Generalized Additive Models (GAMs). Both simulation studies and some performance indicators are used to validate the proposed procedure, showing that, especially in presence of long gaps, spatio-temporal FPCA provides a better reconstruction than spatial FPCA.