Stanczyk / Stanczyk / Zielosko Advances in Feature Selection for Data and Pattern Recognition

1. Auflage 2018
ISBN: 978-3-319-67588-6
Verlag: Springer Nature Switzerland
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, Band 138, 328 Seiten

Reihe: Intelligent Systems Reference Library

This book presents recent developments and research trends in the field of feature selection for data and pattern recognition, highlighting a number of latest advances.

The field of feature selection is evolving constantly, providing numerous new algorithms, new solutions, and new applications. Some of the advances presented focus on theoretical approaches, introducing novel propositions highlighting and discussing properties of objects, and analysing the intricacies of processes and bounds on computational complexity, while others are dedicated to the specific requirements of application domains or the particularities of tasks waiting to be solved or improved.

Divided into four parts – nature and representation of data; ranking and exploration of features; image, shape, motion, and audio detection and recognition; decision support systems, it is of great interest to a large section of researchers including students, professorsand practitioners.

Autoren/Hrsg.

Stanczyk, Urszula

Zielosko, Beata

Jain, Lakhmi C.

Inhaltsverzeichnis

1;Preface;6
2;Contents;8
3;Editors and Contributors;15
4;1 Advances in Feature Selection for Data and Pattern Recognition: An Introduction;19
4.1;1.1 Introduction;19
4.2;1.2 Chapters of the Book;21
4.3;1.3 Concluding Remarks;25
4.4;References;25
5;Part I Nature and Representation of Data;28
6;2 Attribute Selection Based on Reduction of Numerical Attributes During Discretization;29
6.1;2.1 Introduction;29
6.2;2.2 Dominant Attribute Discretization;30
6.3;2.3 Multiple Scanning Discretization;35
6.4;2.4 Experiments;35
6.5;2.5 Conclusions;39
6.6;References;39
7;3 Improving Bagging Ensembles for Class Imbalanced Data by Active Learning;41
7.1;3.1 Introduction;41
7.2;3.2 Improving Classifiers Learned from Imbalanced Data;44
7.2.1;3.2.1 Nature of Imbalanced Data;44
7.2.2;3.2.2 Evaluation of Classifiers on Imbalanced Data;45
7.2.3;3.2.3 Main Approaches to Improve Classifiers for Imbalanced Data;47
7.3;3.3 Active Learning;48
7.4;3.4 Ensembles Specialized for Imbalanced Data;49
7.5;3.5 Active Selection of Examples in Under-Sampling Bagging;51
7.6;3.6 Experimental Evaluation;55
7.6.1;3.6.1 Experimental Setup;55
7.6.2;3.6.2 Results of Experiments;56
7.7;3.7 Conclusions;64
7.8;References;66
8;4 Attribute-Based Decision Graphs and Their Roles in Machine Learning Related Tasks;69
8.1;4.1 Introduction;70
8.2;4.2 The Attribute-Based Decision Graph Structure (AbDG) for Representing Training Data;72
8.2.1;4.2.1 Constructing an AbDG;72
8.2.2;4.2.2 Assigning Weights to Vertices and Edges;74
8.2.3;4.2.3 Computational Complexity for Building an AbDG;77
8.3;4.3 Using the AbDG Structure for Classification Tasks;78
8.4;4.4 Using an AbDG for Classification Purposes - A Case Study;79
8.5;4.5 Using the AbDG Structure for Imputation Tasks;80
8.6;4.6 Searching for Refined AbDG Structures via Genetic Algorithms;82
8.7;4.7 Conclusions;85
8.8;References;86
9;5 Optimization of Decision Rules Relative to Length Based on Modified Dynamic Programming Approach;88
9.1;5.1 Introduction;88
9.2;5.2 Background;90
9.3;5.3 Main Notions;91
9.4;5.4 Modifed Algorithm for Directed Acyclic Graph Construction ??ast(T);93
9.5;5.5 Procedure of Optimization Relative to Length;96
9.6;5.6 Experimental Results;98
9.6.1;5.6.1 Attributes' Values Selection and Size of the Graph;98
9.6.2;5.6.2 Comparison of Length of ?-Decision Rules;100
9.6.3;5.6.3 Classifier Based on Rules Optimized Relative to Length;105
9.7;5.7 Conclusions;106
9.8;References;106
10;Part II Ranking and Exploration of Features;109
11;6 Generational Feature Elimination and Some Other Ranking Feature Selection Methods;110
11.1;6.1 Introduction;111
11.2;6.2 Selected Methods and Algorithms;112
11.2.1;6.2.1 Rough Set Based Feature Selection;113
11.2.2;6.2.2 Random Forest Based Feature Selection;115
11.2.3;6.2.3 Generational Feature Elimination Algorithm;116
11.3;6.3 Feature Selection Experiments;118
11.4;6.4 Results and Conclusions;120
11.5;References;123
12;7 Ranking-Based Rule Classifier Optimisation;126
12.1;7.1 Introduction;126
12.2;7.2 Background;127
12.2.1;7.2.1 Attribute Ranking;128
12.2.2;7.2.2 Rule Classifiers;129
12.2.3;7.2.3 Filtering Rules;130
12.3;7.3 Research Framework;131
12.3.1;7.3.1 Preparation of the Input Datasets;131
12.3.2;7.3.2 Rankings of Features;132
12.3.3;7.3.3 DRSA Decision Rules;133
12.3.4;7.3.4 Weighting Rules;135
12.4;7.4 Experimental Results;136
12.4.1;7.4.1 Filtering Rules by Attributes;136
12.4.2;7.4.2 Filtering Rules by Weights;139
12.4.3;7.4.3 Summary of Results;140
12.5;7.5 Conclusions;142
12.6;References;143
13;8 Attribute Selection in a Dispersed Decision-Making System;145
13.1;8.1 Introduction;145
13.2;8.2 Related Works;146
13.3;8.3 Basics of the Rough Set Theory;147
13.4;8.4 An Overview of Dispersed Systems;148
13.4.1;8.4.1 Basic Definitions;149
13.4.2;8.4.2 Static Structure;150
13.4.3;8.4.3 Dynamic Structure with Disjoint Clusters;151
13.4.4;8.4.4 Dynamic Structure with Inseparable Clusters;153
13.4.5;8.4.5 Dynamic Structure with Negotiations;153
13.5;8.5 Description of the Experiments;155
13.5.1;8.5.1 Data Sets;155
13.5.2;8.5.2 Attribute Selection;156
13.5.3;8.5.3 Evaluation Measures and Parameters Optimization;157
13.6;8.6 Experiments and Discussion;160
13.6.1;8.6.1 Results for Static Structure;160
13.6.2;8.6.2 Results for Dynamic Structure with Disjoint Clusters;169
13.6.3;8.6.3 Results for Dynamic Structure with Inseparable Clusters;169
13.6.4;8.6.4 Results for Dynamic Structure with Negotiations;169
13.6.5;8.6.5 Comparison of All Methods;170
13.6.6;8.6.6 Comparison for Data Sets;172
13.7;8.7 Conclusion;172
13.8;References;173
14;9 Feature Selection Approach for Rule-Based Knowledge Bases;175
14.1;9.1 Introduction;175
14.2;9.2 Feature Selection Methods for Rule Mining Processes;176
14.2.1;9.2.1 Significance of Rules Mining Process;178
14.2.2;9.2.2 Feature Selection in the Rules and Their Clusters;179
14.2.3;9.2.3 Related Works;180
14.2.4;9.2.4 Clustering the Rules Based on the Similarity Approach;180
14.2.5;9.2.5 Clustering the Rules;182
14.2.6;9.2.6 Cluster's Representative;184
14.3;9.3 Rough Set Approach in Creating Rules' Clusters Representatives;185
14.3.1;9.3.1 Lower and Upper Approximation Approach;186
14.3.2;9.3.2 KbExplorer - A Tool for Knowledge Base Exploration;187
14.4;9.4 Experiments;188
14.4.1;9.4.1 From Original Data to the Rules;188
14.4.2;9.4.2 The Results of the Experiments;188
14.4.3;9.4.3 The Summary of the Experiments;192
14.5;9.5 Conclusion;192
14.6;References;193
15;Part III Image, Shape, Motion, and Audio Detection and Recognition;195
16;10 Feature Selection with a Genetic Algorithm for Classification of Brain Imaging Data;196
16.1;10.1 Introduction;197
16.2;10.2 Materials and Methods;200
16.2.1;10.2.1 Data and Preprocessing;200
16.2.2;10.2.2 Least Absolute Shrinkage and Selection Operator;200
16.2.3;10.2.3 Minimalist Genetic Algorithm for Feature Selection;202
16.2.4;10.2.4 Experimental Set-Up;205
16.3;10.3 Results;206
16.4;10.4 Discussion;209
16.5;10.5 Conclusions;210
16.6;References;211
17;11 Shape Descriptions and Classes of Shapes. A Proximal Physical Geometry Approach;214
17.1;11.1 Introduction;215
17.2;11.2 Introduction to Simplicial Complexes;216
17.2.1;11.2.1 Examples: Detecting Shapes from Simplicial Complexes;217
17.3;11.3 Preliminaries of Proximal Physical Geometry;220
17.3.1;11.3.1 Image Object Shape Geometry;220
17.3.2;11.3.2 Descriptions and Proximities;221
17.3.3;11.3.3 Descriptive Proximities;223
17.3.4;11.3.4 Edelsbrunner--Harer Nerve Simplexes;226
17.4;11.4 Features of Image Object Shapes;229
17.5;11.5 Concluding Remarks;234
17.6;References;234
18;12 Comparison of Classification Methods for EEG Signals of Real and Imaginary Motion;237
18.1;12.1 Introduction;238
18.2;12.2 EEG Signal Parameterisation;239
18.3;12.3 Data Classification Method;242
18.4;12.4 Classification Results;244
18.5;12.5 Conclusions;246
18.6;References;247
19;13 Application of Tolerance Near Sets to Audio Signal Classification;250
19.1;13.1 Introduction;250
19.2;13.2 Related Work;252
19.3;13.3 Automatic Detection of Video Blocks;254
19.4;13.4 Audio and Video Features;255
19.4.1;13.4.1 Audio Features;255
19.4.2;13.4.2 Video Features;258
19.5;13.5 Theoretical Framework: Near Sets and Tolerance Near Sets;258
19.5.1;13.5.1 Preliminaries;259
19.5.2;13.5.2 Tolerance Near Sets;261
19.6;13.6 Tolerance Class Learner - TCL;263
19.6.1;13.6.1 Algorithms;264
19.6.2;13.6.2 Phase II: Classification;266
19.7;13.7 Experiments;266
19.7.1;13.7.1 Speech and Non-speech Dataset;266
19.7.2;13.7.2 Discussion;268
19.7.3;13.7.3 Music and Music + Speech Dataset;269
19.7.4;13.7.4 Detection of Commercial Blocks in News Data;270
19.8;13.8 Conclusion and Future Work;272
19.9;References;272
20;Part IV Decision Support Systems;276
21;14 Visual Analysis of Relevant Features in Customer Loyalty Improvement Recommendation;277
21.1;14.1 Introduction;278
21.2;14.2 Related Applications;278
21.3;14.3 Dataset and Application;279
21.3.1;14.3.1 Dataset;279
21.4;14.4 Decision Problem;280
21.4.1;14.4.1 Attribute Analysis;280
21.4.2;14.4.2 Attribute Reduction;281
21.4.3;14.4.3 Customer Satisfaction Analysis and Recognition;281
21.4.4;14.4.4 Providing Recommendations;282
21.5;14.5 Proposed Approach;282
21.5.1;14.5.1 Machine Learning Techniques;282
21.5.2;14.5.2 Visualization Techniques;290
21.6;14.6 Evaluation Results;294
21.6.1;14.6.1 Single Client Data (Local) Analysis;295
21.6.2;14.6.2 Global Customer Sentiment Analysis and Prediction;296
21.6.3;14.6.3 Recommendations for a Single Client;297
21.7;14.7 Conclusions;300
21.8;References;300
22;15 Evolutionary and Aggressive Sampling for Pattern Revelation and Precognition in Building Energy Managing System with Nature-Based Methods for Energy Optimization;302
22.1;15.1 Introduction;303
22.2;15.2 Ant Colony Optimization;303
22.2.1;15.2.1 Ant System;303
22.2.2;15.2.2 Ant Colony Optimization Metaheuristics;305
22.2.3;15.2.3 Ant Colony System;306
22.2.4;15.2.4 Type of Issues That Are Able To Be Solved Using ACO;309
22.3;15.3 Building's Installation as 3D Matrix-Like Grids;310
22.4;15.4 Classical ACO Approach in Revealing a Pattern;313
22.4.1;15.4.1 Capability of Revealing with ACO;313
22.5;15.5 Aggressive ACO Sampling in Revealing a Pattern;315
22.5.1;15.5.1 Ant Decomposition and Specialization;315
22.5.2;15.5.2 Precognition Paths;316
22.5.3;15.5.3 Ants Spreading;318
22.6;15.6 Algorithm and Procedures;319
22.7;15.7 Conclusion;324
22.8;References;325
23;Index;327