Pietro Mantovan

Characterization of rainwater quality from the Venice region network using multiway data analysis

Abstract The paper describes the results of an exploratory analysis of the relations between variables characterizing the quality of atmospheric precipitation in the Veneto, an Italian region with an area of ∼ 18000 km2 and a population of just over 4.5 million. A network of eight sampling stations for monitoring rainfall was set up in 1988, in line with the EMEP statements. The data consisted of 1174 observations over the period February 1989–December 1991. Principal component analysis (PCA) was used in an attempt to describe the structure of relations between the solutes in wet deposition. Some extensions of PCA (interstructure-compromise-intrastructure method) were considered in order to evaluate differences between relation structures as defined by location, season and volume of precipitations.

Sediments and pollution in the Northern Adriatic Sea: a statistical analysis

Abstract Data concerning grain-size distribution, mineralogy, geochemistry, and chlorinated hydrocarbon concentrations, obtained from 246 surface sediment samples collected in the Northern Adriatic sea (Italian area), provide the basis for a multivariate statistical analysis which evaluates the influence of riverine inputs, the differences among various depositional environments and the presence of polluted zones. Cluster analysis applied to grain size has identified nine groups of lithotypes associated with the morphology and the hydrology of the basin. In the Gulf of Venice, principal component analysis on clay minerals and carbonates has permitted recognition of two main areas located north and south of the Brenta river. More subareas correspond to other lithological types. Principal component analysis of nine metal (Hg, Cd, Cu, Pb, Cr, Zn, Fe, Ni, and Co) concentrations reveals that the studied sea area is more complex from the geochemical point of view. Three geochemical zones related to riverine inputs, different lithotypes and specific forms of contamination are identified. High pollution of both halocarbons and heavy metals is associated with fine grain size.

A comparison between parallel algorithms for system parameter estimation in dynamic linear models

When dealing with high-frequency time series, statistical procedures giving reliable estimates of unknown parameters and forecasts in real time are required. This is why recursive estimation methods are usually preferred to maximum-likelihood estimators. In the paper, a recursive estimation algorithm for the system parameter of dynamic linear models is proposed. A comparison with some other algorithms is given via Monte Carlo simulations. Consistency properties of the algorithms are also empirically verified. Copyright

Recursive Estimation of System Parameter in Environmental Time Series Models

Dealing with high-frequency time series, such as environmental ones, raises important inferential and computational problems. Environmental monitoring and forecasting, for instance, require statistical procedures giving reliable estimates of unknown parameters and forecasts in real time. In this paper we consider dynamic linear models as a basic tool for the analysis of such kind of data and propose a recursive estimator for system parameter. A comparison of this estimator with some other estimation methods is provided via Monte Carlo simulations. The estimator we propose is computationally efficient and very easy to implement. Moreover, in our simulation study, it exhibits good asymptotic properties.

Flexible Dynamic Regression Models for Real-time Forecasting of Air Pollutant Concentration

The application of the dynamic regression model to real-time forecasting of air pollutant concentration points out some problems due to both the high frequency of sampling and the need of many-step-ahead forecasting. Some flexible definitions of the system equation are proposed to solve these problems. The proposed definitions are evaluated by means of an application to the prediction of nitrogen dioxide concentration in Venezia-Mestre.

Covariate Error Bias Effects in Dynamic Regression Model Estimation and Improvement in the Prediction by Covariate Local Clusters

We consider a dynamic linear regression model with errors-in-covariate. Neglecting such errors has some undesirable effects on the estimates obtained with the Kalman Filter. We propose a modification of the Kalman Filter where the perturbed covariate is replaced with a suitable function of a local cluster of covariates. Some results of both a simulation experiment and an application are reported.

Variable Architecture Bayesian Neural Networks: Model Selection based on EMC

This work addresses the problem of Selecting appropriate architectures for Bayesian Neural Networks (BNN). Specifically, it proposes a variable architecture model where the number of hidden units are selected by using a variant of the real-coded Evolutionary Monte Carlo algorithm developed by Liang and Wong (2001) for inference and prediction in fixed architecture Bayesian Neural Networks.

System error variance tuning in state-space models

The multivariate dynamic regression model is a particular specification of the dynamic linear model. For this model, we propose a recursive equation for the estimation of the system error variance matrix. The solution can be used when more observation are available at each state of the system. In these cases, the algorithm allows to define a recursive procedure for the estimate of both the state vector (the regression coefficients) and the other hyperparameters of the model. The performances of the proposed method are evaluated by means of Monte Carlo experiments.

Perturbation Models for Principal Component Analysis of Rainwater Pollution Data

The study of the correlation matrix between ion concentration and the principal component analysis on the related variance matrix are widely used to explore the presence of contamination patterns in rainwater. The paper shows that the covariance between ion concentrations is a perturbed measure, and that the total conductivity can be interpreted as the perturbation factor. Then, the paper describes some strategies for measuring and removing the perturbation and how, by removing this effect, correct contamination patterns can be identified. A summary of the results of an application on data measured by the monitoring network of the Veneto region is proposed.