Part I Instance Selection with Similarity-based Methods Introduction to Instance Selection LI Data: Features and Instances 1.2Purpose and Idea of Instance Selection 1.3Software Packages 1.4Datasets Used in this Book Instance Selection in Classification Tasks 2.1Review of Selected Similarity-based Instance Selection Methods 2.2Reducing Computational Complexity 2.3Other Evaluation Models 2.4Comparison of Selected Similarity-based Instance Selection Methods Instance Selection in Regression Tasks 3.1Threshold-based Instance Selection for Regression Tasks 3.2Discretization-based Instance Selection for Regression Tasks 3.3Data Partitioning Weighting Schemes in Instance Selection 4.1Attribute Weighting 4.2Distance Weighting 4.3Diversity Weighting 4.4Outlier Weighting Ensemble Methods in Instance Selection 5.1Bagging of Instance Selection Algorithms 5.2Experimental Evaluation of Instance Selection Bagging Ensembles for Classification Tasks 5.4 Experimental Evaluation of Instance Selection Bagging Ensembles for Regression Tasks Joint Feature and Instance Selection 6.1Feature Filters 6.2Feature Wrappers 6.3Joined Feature and Instance Selection Part II Instance Selection with Evolutionary Methods Introduction to Evolutionary Optimization 7.1Basics of Genetic Algorithms 7.2Fitness Function and Selection 7.3Crossover 7.4Population Size and Initialization 7.5Mutation 7.6Elitism and Steady State Genetic Algorithms 7.7CHC Genetic Algorithms 7.8Cooperative Coevolution 7.9Multi-Objective Evolutionary Algorithms Single-Objective Evolutionary Instance Selection 8.1Encoding 8.2The Objectives and Fitness Function 8.3k-NN as the Inner Evaluation Algorithm Multi-Objective Evolutionary Instance Selection 9.1Encoding 9.2The Objectives 9.3Pareto Front 9.4Instance Selection Process 9.5Choice of the Multi-Objective Genetic Algorithm and its Parameters 9.6.1Problem Description 9.6.2Experimental Evaluation and Discussion 9.7Assessment of Population Initialization Methods 9.7.1Problem Description 9.7.2Experimental Evaluation and Discussion 9.8Tuning k-NN Parameters 9.8.1Problem Description 9.8.2Experimental Evaluation and Discussion 9.9Evaluating Instance Weighting Scheme 9.9.1Problem Description 9.9.2Experimental Evaluation and Discussion 9.10Comparison with Other Methods 9.11Classification Problems Additional Enhancements 10.1Data Space Partitioning 10.2Multiply Fronts to Extend Range and Prevent Over-fitting Optimization of Evolutionary Instance Selection 11.1Optimization of Genetic Algorithms Parameters 11.2.1Population Size and Multi-parent Crossover 11.2.2Fitness Function 11.2.3Shortening Chromosome 11.3Accelerating Calculations of Distance Matrix 11.4Analysis of Computational Time and Complexity Joint Evolutionary Feature and Instance Selection 12.1Sequential Evolutionary Feature and Instance Selection 12.2Features and Instances Encoded in the Same Chromosome 12.3Coevolutionary Feature and Instance Selection Instance Selection for Multi-Output Data 13.1Multi-Objective Evolutionary Instance Selection for Multi-output Regression 13.2Experimental Evaluation 13.3.1Experimental Setup 13.3.2Experimental Results Part III Instance Selection Embedded into Neural Networks Introduction to Neural Networks 14.1Multilayer Perceptron (MLP) 14.2Error Surface 14.3Neural Network Learning Algorithms 14.4.1Backpropagation and Rprop 14.4.2Variable Step Search Algorithm (VSS) Noise Reduction in Neural Network Learning 15.1Noise Reduction With Error Function Modifications 15.2Static Robust Error Measures 15.3.1Least Trimmed Absolute Values (LTA) 15.3.2Iterative Least Median of Squares (ILMedS) 15.3.3Least Mean Log Squares (LMLS) 15.3.4MAE 15.3.5Median Input Function (MIF) 15.3.6MedSum 15.4Dynamic Robust Error Measures 15.4.1 Trapeziod, Exponential, Three-Parabolic and Triangular Error Functions 15.5Experimental Evaluation and Conclusions Joining Embedded and Similarity-based Instance Selection Joint Feature and Instance Selection from Neural Networks 17.1Feature Selection Embedded into Neural Network Learning 17.2Other Methods from Literature 17.3Data Reduction with Boundary Vectors in Neural Networks 17.4Joint Feature and Instance Selection Embedded into Network Learning Special Neural Networks for Data Selection and Rule Extraction 18.1Network Construction and Training 18.2Rule Extraction and Feature Selection for Classification Tasks 18.3Rule Extraction and Feature Selection for Regression Tasks 18.4Instance Selection 18.6.1Decompositional Rule Extraction from Neural Networks 18.6.2Pedagogical Rule Extraction from Neural Networks 18.6.3Hybrid Rule Extraction from Neural Networks