Data Analysis, Machine Learning and Applications - Proceedings of the 31st Annual Conference of the Gesellschaft für Klassifikation e.V., Albert-Ludwigs-Universität Freiburg, March 7-9, 2007

Preface

Contents

Part I Classification

Distance-based Kernels for Real-valued Data

1 Introduction

2 Kernels and similarities defined on real numbers

3 Semantics and applicability

4 Truncated Euclidean similarity

5 Canberra distance-based similarity

6 Kernels defined on real vectors

7 Conclusions

References

Fast Support Vector Machine Classification of Very Large Datasets

1 Introduction

2 Linear SVM trees

3 Non-linear extension

4 Experiments

5 Conclusion

References

Fusion of Multiple Statistical Classifiers

1 Introduction

2 Classifier fusion

3 Diversity of ensemble members

4 Combination rules

5 Open problems

6 Results of experiments

7 Conclusions

References

Calibrating Margin–based Classifier Scores into Polychotomous Probabilities

1 Introduction

2 Reduction to binary problems

3 Coupling probability estimates

4 Dirichlet calibration

5 Comparison

6 Conclusion

References

Classification with Invariant Distance Substitution Kernels

1 Introduction

2 Background

3 Adjustable invariance

4 Positive definiteness

5 Classification experiments

6 Conclusion

References

Applying the Kohonen Self-organizing Map Networks to Select Variables

1 Introduction

2 A proposition to reduce the number of variables

3 Applications and results

4 Conclusions

References

Computer Assisted Classification of Brain Tumors

1 Introduction

2 Algorithms

3 Results

4 Conclusions

References

Model Selection in Mixture Regression Analysis – A Monte Carlo Simulation Study

1 Introduction

2 Model selection in mixture models

3 Simulation design

4 Results summary

5 Key contributions and future research directions

References

Comparison of Local Classification Methods

1 Introduction

2 Local classification methods

3 Simulation study

4 Summary

References

Incorporating Domain Specific Information into Gaia Source Classification

1 Introduction

2 Classification and parametrization

3 Classification results

4 Summary

References

Identification of Noisy Variables for Nonmetric and Symbolic Data in Cluster Analysis

1 Introduction

2 Characteristics of the HINoV method and its modifications

3 Simulation models

4 Discussion on the simulation results

5 Conclusions

References

Part II Clustering

Families of Dendrograms

1 Introduction

2 A brief introduction to

adic geometry

3

adic dendrograms

4 The space of dendrograms

5 Distributions on dendrograms

6 Hidden vertices

7 Conclusions

Acknowledgements

References

Mixture Models in Forward Search Methods for Outlier Detection

1 Introduction

2 The Forward Search

3 Forward Search and Normal Mixture Models: the graphical approach

4 Forward Search and Normal Mixture Models: the inferential approach

5 Concluding remarks and open issues

References

On Multiple Imputation Through Finite Gaussian Mixture Models

1 Introduction

2 Multiple imputation

3 Label switching

4 Simulation study and results

References

Mixture Model Based Group Inference in Fused Genotype and Phenotype Data

1 Introduction

2 Methods

3 Results

4 Discussion

5 Acknowledgements

References

The Noise Component in Model- based Cluster Analysis

1 Introduction

2 Two variations on the noise component

3 Some theory

4 The EM-algorithm

5 Simulations

6 Conclusion

References

An Artificial Life Approach for Semi- supervised Learning

1 Introduction

2 Artificial life

3 Semi-supervised artificial life

4 Semi-Supervised artificial life for cluster analysis

5 Experimental settings and results

6 Discussion

7 Summary

References

Hard and Soft Euclidean Consensus Partitions

1 Introduction

2 Theory

3 Applications

References

Rationale Models for Conceptual Modeling

1 Subjectivism in the modeling process

2 The design rationale approach

3 Classification of rationale fragments

4 Conclusion

References

Measures of Dispersion and Cluster-Trees for Categorical Data

1 Motivation

2 Measures of dispersion

3 Segmentation

References

Information Integration of Partially Labeled Data

1 Introduction

2 Related work

3 Four problem classes

4 Method

5 Evaluation

6 Conclusion

References

Part III Multidimensional Data Analysis

Data Mining of an On-line Survey - A Market Research Application

1 Introduction

2 Data and objectives

3 Methodology and results

4 Conclusions

References

Nonlinear Constrained Principal Component Analysis in the Quality Control Framework

1 Introduction

2 Constrained principal component analysis

3 Nonlinear Constrained Principal Component Analysis

4 Stability analysis

5 Results and interpretation

6 Concluding remarks

References

Non Parametric Control Chart by Multivariate Additive Partial Least Squares via Spline

1 Introduction

2 Multivariate control charts based on projection methods

3 Application: monitoring the painting process of hot-rolled aluminium foils

4 Conclusion

References

Simple Non Symmetrical Correspondence Analysis

1 Introduction

2 Non symmetrical correspondence analysis

3 Simple non symmetrical correspondence analysis

4 Father’s and son’s occupations data

5 Conclusions

References

Factorial Analysis of a Set of Contingency Tables

1 Introduction

2 Methodology

3 Application

4 Discussion

5 Software notes

References

Part IV Analysis of Complex Data

Graph Mining: Repository vs. Canonical Form

1 Introduction

2 Canonical form pruning

3 Repository of processed subgraphs

4 Comparison

5 Experiments

6 Summary

References

Classification and Retrieval of Ancient Watermarks

1 Introduction

2 Feature extraction

3 Results

4 Conclusion

References

Segmentation and Classification of Hyper- Spectral Skin Data

1 Introduction

2 Labelling

3 Classification

4 Results

5 Conclusion

References

FSMTree: An Efficient Algorithm for Mining Frequent Temporal Patterns

1 Introduction

2 Foundations and related work

3 Algorithms FSMSet and FSMTree

4 Performance evaluation and conclusions

References

A Matlab Toolbox for Music Information Retrieval

1 Motivation and approach

2 Feature extraction

3 Data analysis

4 Application to the study of music and emotion

References

A Probabilistic Relational Model for Characterizing Situations in Dynamic Multi- Agent Systems

1 Introduction

2 Framework for modeling and recognizing situations

3 Modeling situations

4 Recognizing situations

5 Evaluation

6 Conclusions and further work

References

Applying the Qn Estimator Online

1 Introduction

2 An update algorithm for the Qn and the HL estimator

3 Comparative study

4 Conclusions

References

A Comparative Study on Polyphonic Musical Time Series Using MCMC Methods

1 Introduction

2 Polyphonic model

3 Extended polyphonic model

4 Results

5 Conclusion

References

Collective Classification for Labeling of Places and Objects in 2D and 3D Range Data

1 Introduction

2 Related work

3 Collective classification

4 Feature extraction in 2D maps

5 Feature selection

6 Experiments

7 Conclusions

8 Acknowledgment

References

Lag or Error? - Detecting the Nature of Spatial Correlation

1 Introduction

2 Model and test statistics

3 Monte Carlo study

4 Results

References

Part V Exploratory Data Analysis and Tools for Data Analysis

Urban Data Mining Using Emergent SOM

1 Introduction

2 Inspection and transformation of data

3 Method

4 Results

5 Conclusion

References

The Konstanz Information Miner

1 Overview

2 Architecture

3 Repository

4 Extending

5 Conclusion

References

A Pattern Based Data Mining Approach

1 Current situation in data mining

2 Introduction to patterns

3 Some data mining patterns

4 Summary and outlook

References

A Framework for Statistical Entity Identification in

1 Introduction

2 Methodological framework

3 Implementation

4 Conclusion and future work

References

Combining Several SOM Approaches in Data Mining: Application to ADSL Customer Behaviours Analysis

1 Introduction

2 Network measurements and data description

3 Customer segmentation

4 Conclusion

References

On the Analysis of Irregular Stock Market Trading Behavior

1 Introduction

2 Irregular trading behavior in a market

3 Analysis of trading behavior with complex valued Eigensystem analysis

4 Analysis of the dataset

5 Conclusion

References

A Procedure to Estimate Relations in a Balanced Scorecard

1 Related work

2 Balanced scorecards

3 Model

4 Case study

5 Results

6 Conclusion and outlook

References

The Application of Taxonomies in the Context of Configurative Reference Modelling

1 Introduction

2 Configurative Reference Modelling and the application of taxonomies

3 Conclusion

4 Outlook

References

Two-Dimensional Centrality of a Social Network

1 Introduction

2 The procedure

3 The analysis and the result

4 Discussion

References

Benchmarking Open-Source Tree Learners in R/RWeka

1 Introduction

2 Design of the benchmark experiment

3 Results of the benchmark experiment

4 Discussion and further work

References

From Spelling Correction to Text Cleaning – Using Context Information

1 Introduction

2 Linguistics and context sensitivity

3 Framework for text preparation

4 Experimental results

5 Conclusion and future work

References

Root Cause Analysis for Quality Management

1 Introduction

2 Root Cause Analysis

3 Computational results

4 Conclusion

References

Finding New Technological Ideas and Inventions with Text Mining and Technique Philosophy

1 Introduction

2 A common structure for raw and context information

3 Relevant aspects for the text mining approach from technique philosophy

4 A text mining approach for

new ideas and inventions

5 Evaluation and outlook

6 Acknowledge

References

Investigating Classifier Learning Behavior with Experiment Databases

1 Introduction

2 A database for classification experiments

3 The experiments

4 Using the database

5 Conclusions

References

Part VI Marketing and Management Science

Conjoint Analysis for Complex Services Using Clusterwise Hierarchical Bayes Procedures

1 Introduction

2 Preference measurement for services

3 Hierarchical Bayes procedures for conjoint analysis

4 Empirical investigation

5 Conclusion and outlook

References

Building an Association Rules Framework for Target Marketing

1 Introduction

2 A segment-specific view of cross-category associations

3 Methodology

4 Empirical application

5 Conclusion and future work

References

AHP versus ACA – An Empirical Comparison

1 Preference measurement for complex products

2 The Analytic Hierarchy Process – AHP

3 Design of the empirical study

4 Results

5 Conclusions and outlook

References

On the Properties of the Rank Based Multivariate Exponentially Weighted Moving Average Control Charts

1 Introduction

2 Data depth

3 The proposed

control chart

4 Effect of the reference sample size on

control charts

performance

5 Conclusion

Acknowledgements

References

Are Critical Incidents Really Critical for a Customer Relationship? A MIMIC Approach

1 Introduction

2 Hypotheses

3 Method

4 Results

5 Discussion

References

Heterogeneity in the Satisfaction-Retention Relationship – A Finite- mixture Approach

1 Introduction

2 The Model

3 Discussion

References

An Early-Warning System to Support Activities in the Management of Customer Equity and How to Obtain the Most from Spatial Customer Equity Potentials

1 Introduction1

2 Strategic customer control dimensions

3 Early-warning system

4 Empirical example

5 Conclusion

References

Classifying Contemporary Marketing Practices

1 Introduction

2 Knowledge on interactive marketing

3 A Finite Mixture approach for classifying marketing practices

4 Empirical application

5 Conclusions

References

Part VII Banking and Finance

Predicting Stock Returns with Bayesian Vector Autoregressive Models

1 Introduction

2 Literature review

3 Model

4 Empirical study

5 Conclusion and outlook

References

The Evaluation of Venture-Backed IPOs – Certification Model versus Adverse Selection Model, Which Does Fit Better?

1 Introduction

2 The theoretical

background: the certification model

and the adverse selection model

3 Data set and non-parametric hypothesis tests

4 Multivariate investigation tools: Partial Least squares regression model

5 Conclusion

Acknowledgments

References

Using Multiple SVM Models for Unbalanced Credit Scoring Data Sets

1 Introduction

2 SVM models for unbalanced data sets

3 Multiple SVM for unbalanced data sets in practice

4 Combination of SVM on random input subsets

5 Conclusions and outlook

References

Part VIII Business Intelligence

Comparison of Recommender System Algorithms Focusing on the New- item and User- bias Problem

1 Introduction

2 Related works

3 Observed approaches

4 Evaluation protocols

5 Evaluation and experimental results

6 Conclusion

References

Collaborative Tag Recommendations

1 Introduction

2 Related work

3 Recommender Systems

4 Tag Recommender Systems

5 Experimental setup and results

6 Conclusions

7 Acknowledgments

References

Applying Small Sample Test Statistics for Behavior- based Recommendations

1 Introduction

2 The ideal decision maker: The decision maker without preferences

3 Library meta catalogs: An exemplary application area

4 Mathematical notation

5 POSICI: Probability Of Single Item Co-Inspections

6 POMICI: Probability Of Multiple Items Co-Inspections

7 POSICI vs. POMICI

8 Conclusions and further research

References

Part IX Text Mining, Web Mining, and the Semantic Web

Classifying Number Expressions in German Corpora

1 Introduction

2 Classification of number expressions

3 Experimental evaluation

4 Conclusions and future work

References

Non-Profit Web Portals - Usage Based Benchmarking for Success Evaluation

1 Introduction

2 Related work

3 Method

4 Case study

5 Conclusions

References

Text Mining of Supreme Administrative Court Jurisdictions

1 Introduction

2 Administrative Supreme Court jurisdictions

3 Investigations

4 Conclusion

References

Supporting Web-based Address Extraction with Unsupervised Tagging

1 Introduction

2 Data preparation

3 Unsupervised tagging

4 Experiments and evaluation

5 Conclusion and further work

References

A Two-Stage Approach for Context-Dependent Hypernym Extraction

1 Introduction

2 Document clustering

3 Hypernym extraction

4 Evaluation

5 Conclusion and future work

References

Analysis of Dwell Times in Web Usage Mining

1 Introduction

2 Model specification and estimation

3 Real life example

4 Conclusion

References

New Issues in Near-duplicate Detection

1 Introduction

2 Fingerprint construction

3 Wikipedia as evaluation corpus

4 Summary

References

Comparing the University of South Florida Homograph Norms with Empirical Corpus Data

1 Introduction

2 Resources

3 Approach

4 Results and discussion

5 Conclusions and future work

Acknowledgments

References

Content-based Dimensionality Reduction for Recommender Systems

1 Introduction

2 Related work

3 The proposed approach

4 Performance study

5 Conclusions

References

Part X Linguistics

The Distribution of Data in Word Lists and its Impact on the Subgrouping of Languages

1 General situation

2 Special situation

3 The bias

4 Solution and operationalization

5 Discussion

6 Conclusions

References

Quantitative Text Analysis Using L-, F- and T- Segments

1 Introduction

2 Data

3 Distribution of segment types

4 Length distribution of L-segments

5 TTR studies

6 Conclusion

References

Projecting Dialect Distances to Geography: Bootstrap Clustering vs. Noisy Clustering

1 Introduction

2 Background and motivation

3 Bootstrapping clustering

4 Clustering with noise

5 Projecting to geography

6 Results

7 Discussion

Acknowledgments

References

Structural Differentiae of Text Types – A Quantitative Model

1 Introduction

2 Category selection

3 The evaluation procedure

4 Exploring the structural homogeneity of text types by means of the Iterative Categorisation Procedure ( ICP)

5 Results

6 Discussion

7 Conclusion

References

Part XI Data Analysis in Humanities

Scenario Evaluation Using Two-mode Clustering Approaches in Higher Education

1 Introduction: Scenario analysis

2 Two-Mode clustering (for scenario evaluation)

3 Example: Scenario evaluation in higher education

4 Conclusions

References

Visualization and Clustering of Tagged Music Data

1 Introduction

2 Related work

3 Emergent Self Organizing Maps

4 Data

5 Experimental results

6 Conclusion and future work

References

Effects of Data Transformation on Cluster Analysis of Archaeometric Data

1 Introduction

2 Data transformation in archaeometry

3 Transformation into ranks

4 Distances and cluster analysis

5 Romano-British vessel glass classified

6 Roman bricks and tiles classified

7 Summary

References

Fuzzy PLS Path Modeling: A New Tool For Handling Sensory Data

1 Introduction

2 Fuzzy PLS path modeling

3 Application

4 Conclusion

References

Automatic Analysis of Dewey Decimal Classification Notations

1 Introduction

2 DDC notations

3 Automatic analysis of DDC notations

4 Results

5 Conclusion

References

A New Interval Data Distance Based on the Wasserstein Metric

1 Introduction

2 A brief survey of the existing distances

3 Our proposal: Wasserstein distance

4 Dynamic clustering algorithm using different criterion functions

5 Conclusion and perspectives

References

Keywords

Author Index