Compstat: Proceedings in Computational Statistics 8th Symposium Held in Copenhagen 1988

ISBN-13: 9783790804119 / Angielski / Miękka / 1988 / 451 str.

David Edwards; Niels E. Raun

Opis

The papers assembled in this volume were presented at COMPSTAT 1988, the 8th biannual Symposium in Computational Statistics held under the auspices of the International Association for Statistical Computing. The current impact of computers on the theory and practice of statistics can be traced at many levels: on one level, the ubiquitous personal computer has made methods for explorative data analysis and display, rarely even described in conventional statistics textbooks, widely available. At another level, advances in computing power permit the development and application of statistical methods in ways that previously have been infeasible. Some of these methods, for example Bayesian methods, are deeply rooted in the philosophical basis of statistics, while others, for example dynamic graphics, present the classical statistical framework with quite novel perspectives. The contents of this volume provide a cross-section of current concerns and interests in computational statistics. A dominating topic is the application of artificial intelligence to statistics (and vice versa), where systems deserving the label" expert systems" are just beginning to emerge from the haze of good intentions with which they hitherto have been clouded. Other topics that are well represented include: nonparametric estimation, graphical techniques, algorithmic developments in all areas, projection pursuit and other computationally intensive methods. COMPSTAT symposia have been held biannually since 1974. This tradition has made COMPSTAT a major forum for advances in computational statistics with contributions from many countries in the world. Two new features have been introduced at COMPSTAT '88.

Spis treści

Keynote Papers.- Parallel Linear Algebra in Statistical Computations.- Three Examples of Computer-Intensive Statistical Inference.- Non-Parametric Estimation.- Efficient Nonparametric Smoothing in High Dimensions Using Interactive Graphical Techniques.- A Boundary Modification of Kernel Function Smoothing, with Application to Insulin Absorption Kinetics.- A Roughness Penalty Regression Approach for Statistical Graphics.- Projection Pursuit.- Detecting Structures by Means of Projection Pursuit.- Confidence Regions for Projection Pursuit Density Estimates.- A Robustness Property of the Projection Pursuit Methods in Sampling from Separably Dependent Random Vectors.- Graphical Techniques.- Graphical Modelling with Large Numbers of Variables: An Application of Principal Components.- Some Graphical Displays for Square Tables.- Data Plotting Methods for Checking Multivariate Normality and Related Ideas.- Computer-aided Illustration of Regression Diagnostics.- Computer Guided Diagnostics.- Expert Systems.- How should the Statistical Expert System and its User see Each Other?.- Towards a Probabilistic Analysis of MYCIN-like Expert Systems.- An Expert System Accepting Knowledge in a form of Statistical Data.- Building a Statistical Expert System with Knowledge Bases of Different Levels of Abstraction.- An Expert System for the Interpretation of Results of Canonical Covariance Analysis.- Building a Statistical Knowledge Base: A Discussion of the Approach used in the Development of THESEUS, a Statistical Expert System.- Prince: An Expert System for Nonlinear Principal Components Analysis.- Expert Systems for Non-Linear Modelling: Progress and Prospects.- Inside a Statistical Expert System: Statistical Methods Employed in the ESTES System.- An Implementation of an EDA Expert System in Prolog Environment.- Automatic Acquisition of Knowledge Base from Data without Expert: ESOD (Expert Sytem from Observational Data).- Experiments with Probabilistic Consultation Systems.- Statistical Consultants and Statistical Expert Systems.- On Inference Process.- Identification Keys, Diagnostic Tables and Expert Systems.- Languages and Packages.- Adding new Statistical Techniques to Standard Software Systems: A Review.- Funigirls: A Prototype Functional Programming Language for the Analysis of Generalized Linear Models.- Blinwdr: An APL-function Library for Interactively Solving the Problem of Robust and Bounded Influence Regression.- Computational Intensive Methods.- Exact Non-Parametric Significance Tests.- Resampling Tests of Statistical Hypotheses.- Clustering Based on Neural Network Processing.- Decision Tree Classifier for Speech Recognition.- Algorithms.- Efficient Sampling Algorithms and Balanced Samples.- Recursive Partition in Biostatistics: Stability of Trees and Choice of the Most Stable Classification.- Generating Rules by Means of Regression Analysis.- A New Algorithm for Matched Case-Control Studies with Applications to Additive Models.- An Algorithm for the Approximation of N-Dimensional Distributions.- Further Recursive Algorithms for Multidimensional Table Computation.- Statistical Methods.- Nonlinear Regression: Methodological and Software Aspects.- Comparing Sensitivity of Models to Missing Data in the GMANOVA.- A Modelling Approach to Multiple Correspondence Analysis.- Multidimensional Scaling on the Sphere.- A Monte Carlo Evaluation of the Methods for Estimating the Parameters of the Generalized Lambda Distribution.- Statistical Guidance for Model Modification in Covariance Structure Analysis.- Similarities Functions.- Robust Bayesian Regression Analysis with HPD-Regions.- Time Series.- Estimation of ARMA Process Parameters and Noise Variance by Means of a Non-Linear Filtering Algorithm.- Autoregressive Models with Latent Variables.- An Algorithm for Time Series Decomposition Using StateSpace Models with Singular Transition Matrix.- Statistical Data Bases and Survey Processing.- New Perspectives in Computer Assisted Survey Processing.- Multiple Imputation for Data-Base Construction.- Grasp: A Complete Graphical Conceptual Language for Definition and Manipulation of Statistical Databases.- Experimental Design.- New Algorithmic and Software Tools for D-Optimal Design Computation in Nonlinear Regression.- Econometric Computing.- Model-Building on Micro Computers: Spreadsheets or Specific Software.- Late Arrivals.- Three Examples of Computer-Intensive Statistical Inference.- New Computer Procedures for Generating Optimal Mixture Designs on Finite Design Spaces.- Screening Based Exclusively on Experts Opinions.- Address list of Authors.

O autorze

David Edwards researches Statistical Bioinformatics at Aarhus University's Department of Molecular Biology and Genetics.