Palumbo / Greenacre Data Analysis and Classification

Proceedings of the 6th Conference of the Classification and Data Analysis Group of the Società Italiana di Statistica
1. Auflage 2010
ISBN: 978-3-642-03739-9
Verlag: Springer
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 482 Seiten

Autoren/Hrsg.

Palumbo, Francesco

Greenacre, Michael

Inhaltsverzeichnis

1;Preface;5
2;List of Referees;8
3;Contents;9
4;Contributors;15
5;Part I Key-note;21
5.1;Clustering of High-Dimensional and Correlated Data;22
5.1.1;1 Introduction;22
5.1.2;2 Definition of Mixture Models;23
5.1.3;3 Maximum Likelihood Estimation;24
5.1.4;4 Choice of Starting Values for the EM Algorithm;24
5.1.5;5 Clustering via Normal Mixtures;25
5.1.6;6 Factor Analysis Model for Dimension Reduction;26
5.1.7;7 Some Recent Extensions for High-Dimensional Data;27
5.1.8;8 Mixtures of Normal Components with Random Effects;28
5.1.9;References;30
5.2;Statistical Methods for Cryptography;31
5.2.1;1 Introduction;31
5.2.1.1;1.1 Different disciplines in cryptography;32
5.2.2;2 Prime Numbers;33
5.2.2.1;2.1 Tests of primality;33
5.2.2.2;2.2 Deterministic tests;34
5.2.2.3;2.3 Some deterministic tests;35
5.2.3;3 The Sum Modulo m of Statistical Variables;36
5.2.4;References;39
6;Part II Cluster Analysis;40
6.1;An Algorithm for Earthquakes Clustering Based on MaximumLikelihood;41
6.1.1;1 Introduction;41
6.1.2;2 Conditional Intensity Function of the Clustering Procedure;42
6.1.2.1;2.1 The ETAS Model;43
6.1.2.2;2.2 Intensity Function for a Particular ClusteredInhomogeneous Poisson Process;44
6.1.3;3 The Proposed Clustering Method;45
6.1.3.1;3.1 Finding a Candidate Cluster and Likelihood Changes;45
6.1.3.2;3.2 The Algorithm of Clustering;46
6.1.4;4 Application to a Real Catalog and Final Remarks;47
6.1.5;References;48
6.2;A Two-Step Iterative Procedure for Clustering of Binary Sequences;49
6.2.1;1 Introduction;49
6.2.2;2 Clustering and Dimensionality Reduction;50
6.2.3;3 Example;53
6.2.4;References;56
6.3;Clustering Linear Models Using Wasserstein Distance;57
6.3.1;1 Introduction;57
6.3.2;2 Input Data and the Clustering Algorithm;58
6.3.3;3 Dynamic Clustering of Linear Models;59
6.3.3.1;3.1 Wasserstein Distance for Distributions ;60
6.3.3.2;3.2 Representation and Allocation Functions;61
6.3.4;4 An Application Using Bank of Italy Household Survey Data;62
6.3.5;5 Conclusions and Perspectives;64
6.3.6;References;64
6.4;Comparing Approaches for Clustering Mixed Mode Data: An Application in Marketing Research;65
6.4.1;1 Introduction;65
6.4.2;2 Obtaining Partitions with Mixed Mode Data;66
6.4.3;3 Some Illustrative Applications;67
6.4.4;4 Discussion;72
6.4.5;References;72
6.5;The Progressive Single Linkage Algorithm Based on Minkowski Ultrametrics;74
6.5.1;1 Introduction;74
6.5.2;2 Ultrametric Approximations;75
6.5.3;3 The Progressive Single Linkage Algorithm;76
6.5.4;4 Some Applications to Real Data;77
6.5.5;5 Conclusions;79
6.5.6;Appendix;80
6.5.7;References;81
6.6;Visualization of Model-Based Clustering Structures;82
6.6.1;1 Introduction;82
6.6.2;2 Dimension Reduction for Model-Based Clustering;83
6.6.3;3 Visualization of Clustering Structures;84
6.6.4;4 Examples;86
6.6.4.1;4.1 Overlapping Clusters with Unconstrained Covariances;86
6.6.4.2;4.2 High Dimensional Mixture of Two Normal Distributions;87
6.6.4.3;4.3 Wisconsin Diagnostic Breast Cancer Data;87
6.6.5;5 Comments and Extensions;89
6.6.6;References;89
7;Part III Multidimensional Scaling;91
7.1;Models for Asymmetry in Proximity Data;92
7.1.1;1 Introduction;92
7.1.2;2 A Class of Scalar Product Models;92
7.1.3;3 The Analysis of Skew-Symmetry with External Information;94
7.1.4;4 Conclusions;96
7.1.5;References;97
7.2;Intimate Femicide in Italy: A Model to Classify How KillingsHappened;98
7.2.1;1 Introduction;98
7.2.2;2 National and International Scenario;99
7.2.3;3 Data and Method;101
7.2.4;4 The Main Results;102
7.2.5;References;104
7.3;Two-Dimensional Centrality of Asymmetric Social Network;105
7.3.1;1 Introduction;105
7.3.2;2 The Procedure;106
7.3.3;3 The Data;107
7.3.4;4 The Analysis;107
7.3.5;5 Result;108
7.3.6;6 Discussion;109
7.3.7;References;111
7.4;The Forward Search for Classical Multidimensional Scaling When the Starting Data Matrix Is Known;113
7.4.1;1 Introduction;113
7.4.2;2 Classical Multidimensional Scaling and the Forward Search;114
7.4.3;3 A Case Study: Linosa Dataset;116
7.4.4;4 Conclusions;120
7.4.5;References;121
8;Part IV Multivariate Analysis and Application;122
8.1;Discriminant Analysis on Mixed Predictors;123
8.1.1;1 Introduction;123
8.1.2;2 Mixed Discriminant Predictors;124
8.1.3;3 Application Example;126
8.1.3.1;3.1 Predictor Analysis;126
8.1.3.2;3.2 Discriminant Analysis;128
8.1.3.3;3.3 Comparison;128
8.1.4;4 Conclusion;130
8.1.5;References;130
8.2;A Statistical Calibration Model for Affymetrix Probe Level Data;131
8.2.1;1 Introduction;131
8.2.2;2 A Brief Review on preprocessing Methods;132
8.2.3;3 The Proposed Calibration Method;133
8.2.4;4 A Comparison with the Most Popular Methods;135
8.2.5;5 Conclusions;137
8.2.6;References;138
8.3;A Proposal to Fuzzify Categorical Variables in Operational Risk Management;139
8.3.1;1 Fuzzy Approach;139
8.3.2;2 The Problem;140
8.3.3;3 The Proposal;141
8.3.4;4 Results;143
8.3.5;5 Conclusions;144
8.3.6;References;145
8.4;Common Optimal Scaling for Customer Satisfaction Models: A Point to Cobb–Douglas' Form;146
8.4.1;1 Features of a Customer Satisfaction model;146
8.4.2;2 Categorical Regression with Common Optimal Scaling;147
8.4.2.1;2.1 The Pattern of the Model;148
8.4.2.1.1;2.1.1 The Algorithm of the Parameters Estimation;149
8.4.3;3 Multiplicative Models for CS;149
8.4.3.1;3.1 Some Observations;150
8.4.4;4 A Theory About Overall CS;151
8.4.5;5 Conclusions;152
8.4.6;References;152
8.5;Structural Neural Networks for Modeling Customer Satisfaction;154
8.5.1;1 Introduction;154
8.5.2;2 Customer Satisfaction and PLS Path Modeling;155
8.5.3;3 Customer Satisfaction and Neural Networks;156
8.5.4;4 A Structural Neural Network for Modeling CS;157
8.5.5;5 Concluding Remarks;161
8.5.6;References;161
8.6;Dimensionality of Scores Obtained witha Paired-Comparison Tournament Systemof Questionnaire Items;163
8.6.1;1 Preference Data Collection;163
8.6.2;2 Preference Data;164
8.6.3;3 Scoring Algorithms;165
8.6.4;4 Dimensions in Preference Data;166
8.6.5;5 Conclusions;170
8.6.6;References;170
8.7;Using Rasch Measurement to Assess the Roleof the Traditional Family in Italy;171
8.7.1;1 Introduction;171
8.7.2;2 Data and Descriptive Analysis;172
8.7.3;3 Method;173
8.7.4;4 Discussion;175
8.7.5;References;177
8.8;Preserving the Clustering Structure by a Projection PursuitApproach;178
8.8.1;1 Introduction;178
8.8.2;2 Projection Pursuit for Preserving the Clustering Structure;180
8.8.2.1;2.1 The Critical Bandwidth to Test Multimodality;180
8.8.2.2;2.2 Projection Pursuit Using the Adjusted Critical Bandwidth;182
8.8.3;3 Numerical Results;182
8.8.3.1;3.1 A Simulation Study;182
8.8.3.2;3.2 Real Data Applications;183
8.8.4;4 Concluding Remarks;185
8.8.5;References;185
8.9;Association Rule Mining of Multimedia Content;186
8.9.1;1 Introduction;186
8.9.2;2 Syntactic Analysis of Video Data;187
8.9.3;3 Semantic Analysis;189
8.9.4;4 Association Rules;190
8.9.5;5 Synergy Effects;191
8.9.6;References;192
9;Part V Classification and Classification Tree;194
9.1;Automatic Dictionary- and Rule-Based Systems for Extracting Information from Text;195
9.1.1;1 Introduction;195
9.1.2;2 A Model for Creating a Meta-dictionary by Means of a Hybrid System;197
9.1.3;3 Application to the Istat TUS Survey;199
9.1.4;References;204
9.2;Several Computational Studies About Variable Selection for Probabilistic Bayesian Classifiers;205
9.2.1;1 Introduction;205
9.2.2;2 Bayesian Networks and Classification;206
9.2.2.1;2.1 Learning Bayesian Networks;207
9.2.2.2;2.2 Classifiers Based on Bayesian Networks;207
9.2.3;3 Feature Subset Selection for Classification;207
9.2.4;4 Experimental Results and Conclusions;208
9.2.5;References;212
9.3;Semantic Classification and Co-occurrences: A Methodfor the Rules Production for the Information Extractionfrom Textual Data;214
9.3.1;1 Introduction;214
9.3.2;2 The Analyzed Corpus and Semantic Classification;215
9.3.3;3 Rules Production Using Co-occurrences and Collocations;215
9.3.4;4 Future Developments and Improvements;220
9.3.5;References;221
9.4;The Effectiveness of University Education: A Structural Equation Model;222
9.4.1;1 Introduction;222
9.4.2;2 The Model;223
9.4.3;3 Main Results;226
9.4.4;4 Conclusions;228
9.4.5;References;229
9.5;Simultaneous Threshold Interaction Detection in BinaryClassification;230
9.5.1;1 Introduction;230
9.5.2;2 Modeling Interaction Effects in Regression Analysis;231
9.5.3;3 The Trunk Model;233
9.5.4;4 Empirical Evidence;234
9.5.5;5 Concluding Remarks;236
9.5.6;References;237
9.6;Detecting Subset of Classifiers for Multi-attribute ResponsePrediction;238
9.6.1;1 Introduction;238
9.6.2;2 The SASSC Algorithm;239
9.6.3;3 Analyzing the Letter Recognition Dataset;242
9.6.4;4 Concluding Remarks;244
9.6.5;References;245
9.7;Clustering Textual Data by Latent Dirichlet Allocation: Applications and Extensions to Hierarchical Data;246
9.7.1;1 Introduction;246
9.7.2;2 The LDA Model;247
9.7.3;3 Considering Documents Structure: An Extension of the LDA Model;249
9.7.4;4 Generalizing the Prior Enrichment: Collapsed Variational Bayes;251
9.7.5;5 Discussion;253
9.7.6;References;253
9.8;Multilevel Latent Class Models for Evaluation of Long-term Care Facilities;254
9.8.1;1 Introduction and Data;254
9.8.2;2 The Model;255
9.8.3;3 The Results;257
9.8.4;4 Conclusions;261
9.8.5;References;261
9.9;Author–Coauthor Social Networks and Emerging ScientificSubfields;262
9.9.1;1 Introduction;262
9.9.2;2 Distribution of Tie Strength;264
9.9.3;3 Distribution of Clique Size;266
9.9.4;4 Random Graph Model for Preferential Attachment;267
9.9.5;5 The Emergence of Scientific Subfields;268
9.9.6;6 The Network of Well-Established Scholars;270
9.9.7;7 Conclusions;272
9.9.8;References;273
10;Part VI Statistical Models;274
10.1;A Hierarchical Model for Time Dependent Multivariate Longitudinal Data;275
10.1.1;1 Introduction;275
10.1.2;2 Model-Based Approach to Three-Way Data Clustering;276
10.1.3;3 Multivariate Hidden Markov Model for Three-Way Data Clustering;277
10.1.4;4 Computational Details;279
10.1.5;5 Simulation Results;280
10.1.6;6 Conclusion;282
10.1.7;References;283
10.2;Covariate Error Bias Effects in Dynamic Regression Model Estimation and Improvement in the Prediction by Covariate Local Clusters;284
10.2.1;1 Introduction;284
10.2.2;2 Bias Effects in Dynamic Regression Models with Errors in the Covariate;285
10.2.3;3 A Local Cluster Kalman Filter;286
10.2.4;4 Simulation Experiments;287
10.2.5;5 An Application;289
10.2.6;6 Conclusions;291
10.2.7;References;291
10.3;Local Multilevel Modeling for Comparisons of InstitutionalPerformance;292
10.3.1;1 Introduction;292
10.3.2;2 Capturing Local Behaviour;293
10.3.2.1;2.1 Mixture Modeling;293
10.3.2.2;2.2 Cluster-Weighted Modeling;294
10.3.3;3 The Proposal: Local Multilevel Modeling;295
10.3.4;4 An Example;296
10.3.5;5 Conclusions and Further Research;298
10.3.6;References;299
10.4;Modelling Network Data: An Introduction to Exponential Random Graph Models;300
10.4.1;1 Introduction;300
10.4.2;2 Exponential Random Graph Models for Social Networks;301
10.4.3;3 The Collaboration Network of Italian Scholars on Population Studies;303
10.4.4;4 Model Estimation Results;305
10.4.5;5 Concluding Remarks;307
10.4.6;References;307
11;Part VII Latent Variables;309
11.1;An Analysis of Earthquakes Clustering Based on a Second-Order Diagnostic Approach;310
11.1.1;1 Introduction;310
11.1.2;2 Second-Order Residual Analysis;311
11.1.2.1;2.1 The Weighted Process and Its Second-Order Properties;311
11.1.2.1.1;2.1.1 The Weighted Spectrum;312
11.1.2.1.2;2.1.2 The Weighted Correlation Integral;312
11.1.3;3 Space–Time ETAS Model;313
11.1.4;4 Nonparametric Estimation and Diagnostics;314
11.1.5;5 Conclusion;317
11.1.6;References;317
11.2;Latent Regression in Rasch Framework;319
11.2.1;1 Introduction;319
11.2.2;2 Longitudinal Latent Regression Model;320
11.2.3;3 Latent Regression Rasch Model with Missing Data;322
11.2.4;4 Conclusion;325
11.2.5;References;326
11.3;A Multilevel Latent Variable Model for Multidimensional Longitudinal Data;328
11.3.1;1 Introduction;328
11.3.2;2 Model for Continuous Responses;329
11.3.2.1;2.1 Model Specification;329
11.3.2.2;2.2 Estimation;331
11.3.2.3;2.3 Application to a Real Data Set;333
11.3.3;3 Conclusion;335
11.3.4;References;335
11.4;Turning Point Detection Using Markov Switching Models with Latent Information;336
11.4.1;1 Introduction;336
11.4.2;2 The Generalized Hamilton Model with Latent Information;337
11.4.3;3 Identifying and Forecasting USA and Japanese Turning Points;339
11.4.4;References;343
12;Part VIII Knowledge Extraction from Temporal Data;344
12.1;Statistical and Numerical Algorithms for Time SeriesClassification;345
12.1.1;1 Introduction;345
12.1.2;2 A Simulation Experiment;348
12.1.3;3 Application to Real Data;350
12.1.4;4 Concluding Remarks;351
12.1.5;References;351
12.2;Mining Time Series Data: A Selective Survey;353
12.2.1;1 Introduction;353
12.2.2;2 Comparing Time Series Shape;354
12.2.3;3 Criteria Based on Fourier and Wavelet Analysis;355
12.2.4;4 Structural Dissimilarity;357
12.2.5;5 Final Remarks;358
12.2.6;References;359
12.3;Predictive Dynamic Models for SMEs;361
12.3.1;1 Introduction;361
12.3.2;2 Default Estimation: A Methodological Proposal;362
12.3.3;3 Data Sources;364
12.3.4;4 Application;364
12.3.5;5 Conclusion;365
12.3.6;References;366
12.4;Clustering Algorithms for Large Temporal Data Sets;367
12.4.1;1 The Framework: Temporal Data Mining;367
12.4.2;2 Temporal Cluster Analysis;368
12.4.3;3 Clustering Algorithms: Applicability to Large Temporal Datasets;370
12.4.4;4 An Example of Application on a Radar Satellite Data Base;372
12.4.5;5 Concluding Remarks;373
12.4.6;References;374
13;Part IX Outlier Detection and Robust Methods;376
13.1;Robust Clustering for Performance Evaluation;377
13.1.1;1 Introduction;377
13.1.2;2 Mahalanobis Distances and the Forward Search;378
13.1.3;3 Example;379
13.1.4;References;385
13.2;Outliers Detection Strategy for a Curve Clustering Algorithm;387
13.2.1;1 Introduction;387
13.2.2;2 Dynamical Curves Clustering with Free knots Spline Estimation;388
13.2.3;3 An Improvement of DCC&FSE Algorithm;389
13.2.4;4 The Outliers Selection Process;390
13.2.5;5 Main Results;391
13.2.6;6 Conclusion and Future Work;394
13.2.7;References;394
13.3;Robust Fuzzy Classification;395
13.3.1;1 Introduction;395
13.3.2;2 Robust Fuzzy Cluster Analysis;396
13.3.3;3 Confirmatory Analysis;398
13.3.4;References;402
13.4;Weighted Likelihood Inference for a Mixed Regressive Spatial Autoregressive Model;403
13.4.1;1 Introduction;403
13.4.2;2 A Weighted Likelihood Approach;404
13.4.3;3 A Small Simulation Study;406
13.4.4;4 A Real Example;408
13.4.5;5 Final Remarks;409
13.4.6;References;410
13.5;Detecting Price Outliers in European Trade Data with the Forward Search;411
13.5.1;1 Introduction;411
13.5.2;2 Application Context, Data and Statistical Patterns;412
13.5.3;3 Application of the Forward Search;413
13.5.4;4 Heuristic Comparison with the ``Backward'' Outliers;416
13.5.5;5 Towards an Automatic Procedure;416
13.5.6;6 Discussion and Main Conclusions;418
13.5.7;References;418
14;Part X Statistical Methods for Financial and Economics Data;420
14.1;Comparing Continuous Treatment Matching Methods in PolicyEvaluation;421
14.1.1;1 Introduction;421
14.1.2;2 Simulating a Subsidies Allocation Mechanism:The Case of L.488;422
14.1.3;3 The Matching Methods in the Continuous Framework;423
14.1.4;4 The Monte Carlo Experiment;424
14.1.5;5 Results and Conclusions;426
14.1.6;References;428
14.2;Temporal Aggregation and Closure of VARMA Models: Some New Results;429
14.2.1;1 Introduction;429
14.2.2;2 Temporal Aggregation;430
14.2.3;3 Temporal Aggregation and VARMA Models;431
14.2.4;4 ``Markovian" Representation;433
14.2.5;References;436
14.3;An Index for Ranking Financial Portfolios According to Internal Turnover;438
14.3.1;1 Introduction;438
14.3.2;2 Style Analysis Models;439
14.3.3;3 Ranking Portfolios According to Internal Turnover;441
14.3.4;4 Concluding Remarks;444
14.3.5;References;445
14.4;Bayesian Hidden Markov Modelsfor Financial Data;446
14.4.1;1 Introduction;446
14.4.2;2 The Model;447
14.4.3;3 Computational Implementation;448
14.4.4;4 Bayesian Inference and Forecasting;450
14.4.5;5 An Application to Financial Data;451
14.4.6;6 Conclusions;453
14.4.7;References;453
15;Part XI Missing Values;455
15.1;Regression Imputation for Space-Time Datasets with MissingValues;456
15.1.1;1 Introduction;456
15.1.2;2 A New Regression Single Imputation Method;458
15.1.3;3 Missing Data Simulation;459
15.1.4;4 Results;461
15.1.5;5 Toward a Multiple Imputation Method;461
15.1.6;References;463
15.2;A Multiple Imputation Approach in a Survey on University Teaching Evaluation;464
15.2.1;1 Introduction;464
15.2.2;2 An Imputation Procedure to Recover for Missingness;465
15.2.3;3 An Application to the Data on the Evaluationof University Teaching;466
15.2.3.1;3.1 AD;468
15.2.3.2;3.2 AE;468
15.2.4;4 Some Final Remarks;472
15.2.5;References;472