COMPSTAT 2008: Proceedings in Computational Statistics

The 18th Conference of IASC-ERS, COMPSTAT’2008,is held in Porto,P- tugal,fromAugust24thtoAugust29th2008,locallyorganisedbytheFaculty of Economics of the University of Porto. COMPSTAT is an initiative of the European Regional Section of the Int- national Association for Statistical Computing (IASC-ERS), a section of the International Statistical Institute (ISI). COMPSTAT conferences started in 1974 in Wien; previous editions of COMPSTAT were held in Berlin (2002), Prague (2004) and Rome (2006). It is one of the most prestigious world conferences in Computational Statistics, regularly attracting hundreds of - searchers and practitioners, and has gained a reputation as an ideal forum for presenting top qualitytheoretical and applied work,promoting interdis- plinary researchand establishing contacts amongstresearcherswith common interests. COMPSTAT’2008 is the ?rst edition of COMPSTAT to be hosted by a Portuguese institution. Keynote lectures are addressed by Peter Hall (Department of Mathematics and Statistics, The University of Melbourne), Heikki Mannila (Department of Computer Science, Faculty of Science, University of Helsinki) and Timo Ter¨ asvirta (School of Economics and Management, University of Aarhus). The conference program includes two tutorials: “Computational Methods in Finance”byJamesGentle(DepartmentofComputationalandDataSciences, George Mason University) and “Writing R Packages” by Friedrich Leisch (Institut fur ¨ Statistik, Ludwig-Maximilians-Universit¨ at). Each COMPSTAT meeting is organised with a number of topics highlighted, which lead to - vited Sessions. The Conference program includes also contributed sessions in di?erent topics (both oral communications and posters).

Keynote.- Nonparametric Methods for Estimating Periodic Functions, with Applications in Astronomy.- Advances on Statistical Computing Environments.- Back to the Future: Lisp as a Base for a Statistical Computing System.- Computable Statistical Research and Practice.- Implicit and Explicit Parallel Computing in R.- Classification and Clustering of Complex Data.- Probabilistic Modeling for Symbolic Data.- Monothetic Divisive Clustering with Geographical Constraints.- Comparing Histogram Data Using a Mahalanobis–Wasserstein Distance.- Computation for Graphical Models and Bayes Nets.- Iterative Conditional Fitting for Discrete Chain Graph Models.- Graphical Models for Sparse Data: Graphical Gaussian Models with Vertex and Edge Symmetries.- Parameterization and Fitting of a Class of Discrete Graphical Models.- Computational Econometrics.- Exploring the Bootstrap Discrepancy.- On Diagnostic Checking Time Series Models with Portmanteau Test Statistics Based on Generalized Inverses and.- New Developments in Latent Variable Models: Non-linear and Dynamic Models.- Computational Statistics and Data Mining Methods for Alcohol Studies.- Estimating Spatiotemporal Effects for Ecological Alcohol Systems.- A Directed Graph Model of Ecological Alcohol Systems Incorporating Spatiotemporal Effects.- Spatial and Computational Models of Alcohol Use and Problems.- Finance and Insurance.- Optimal Investment for an Insurer with Multiple Risky Assets Under Mean-Variance Criterion.- Inhomogeneous Jump-GARCH Models with Applications in Financial Time Series Analysis.- The Classical Risk Model with Constant Interest and Threshold Strategy.- Estimation of Structural Parameters in Crossed Classification Credibility Model Using Linear Mixed Models.- Information Retrieval for Text and Images.- A Hybrid Approach for Taxonomy Learning from Text.- Image and Image-Set Modeling Using a Mixture Model.- Strategies in Identifying Issues Addressed in Legal Reports.- Knowledge Extraction by Models.- Sequential Automatic Search of a Subset of Classifiers in Multiclass Learning.- Possibilistic PLS Path Modeling: A New Approach to the Multigroup Comparison.- Models for Understanding Versus Models for Prediction.- Posterior Prediction Modelling of Optimal Trees.- Model Selection Algorithms.- Selecting Models Focussing on the Modeller’s Purpose.- A Regression Subset-Selection Strategy for Fat-Structure Data.- Fast Robust Variable Selection.- Models for Latent Class Detection.- Latent Classes of Objects and Variable Selection.- Modelling Background Noise in Finite Mixtures of Generalized Linear Regression Models.- Clustering via Mixture Regression Models with Random Effects.- Multiple Testing Procedures.- Testing Effects in ANOVA Experiments: Direct Combination of All Pair-Wise Comparisons Using Constrained Synchronized Permutations.- Multiple Comparison Procedures in Linear Models.- Inference for the Top-k Rank List Problem.- Random Search Algorithms.- Monitoring Random Start Forward Searches for Multivariate Data.- Generalized Differential Evolution for General Non-Linear Optimization.- Statistical Properties of Differential Evolution and Related Random Search Algorithms.- Robust Statistics.- Robust Estimation of the Vector Autoregressive Model by a Least Trimmed Squares Procedure.- The Choice of the Initial Estimate for Computing MM-Estimates.- Metropolis Versus Simulated Annealing and the Black-Box-Complexity of Optimization Problems.- Signal Extraction and Filtering.- Filters for Short Nonstationary Sequences: The Analysis of the Business Cycle.- Estimation of Common Factors Under Cross-Sectional and Temporal Aggregation Constraints: Nowcasting Monthly GDP and Its Main Components.