In memory of Alexey Chervonenkis
Preface
Organization
Contents
Invited Papers
Learning with Intelligent Teacher: Similarity Control and Knowledge Transfer
1 Introduction
2 Learning with Intelligent Teacher: Privileged Information
3 Statistical Analysis of the Rate of Convergence
4 SVM+ for Similarity Control in LUPI Paradigm
5 Three Examples of Similarity Control Using Privileged Information
5.1 Advanced Technical Model as Privileged Information
5.2 Future Events as Privileged Information
5.3 Holistic Description as Privileged Information
6 Transfer of Knowledge Obtained in Privileged Information Space to Decision Space
6.1 Knowledge Representation
6.2 Scheme of Knowledge Transfer Between Spaces
Finding Fundamental Elements of Knowledge.
Fundamental Elements of Knowledge for Homogenous Quadratic Kernel.
Finding Images of Frames in Space X.
6.3 Algorithms for Knowledge Transfer
6.4 Kernels Involved in Intelligent Learning
6.5 Knowledge Transfer for Statistical Inference Problems
6.6 General Remarks About Knowledge Transfer
What Knowledge Does Teacher Transfer?
What Are the Roots of Intelligence?
Holistic Description and Culture.
Quadratic Kernel.
Some Philosophical Interpretations.
7 Conclusions
References
Statistical Inference Problems and Their Rigorous Solutions
1 Basic Concepts of Classical Statistics
1.1 Cumulative Distribution Function
1.2 General Problems of Probability Theory and Statistics
1.3 Empirical Cumulative Distribution Functions
1.4 The Glivenko-Cantelli Theorem and Kolmogorov Type Bounds
1.5 Generalization to Multidimensional Case
2 Main Problems of Statistical Inference
2.1 Conditional Density, Conditional Probability, Regression, and Density Ratio Functions
2.2 Direct Constructive Setting for Conditional Density Estimation
2.3 Direct Constructive Setting for Conditional Probability Estimation
2.4 Direct Constructive Setting for Regression Estimation
2.5 Direct Constructive Setting of Density Ratio Estimation Problem
3 Solution of Ill-Posed Operator Equations
4 Stochastic Ill-Posed Problems
5 Solving Statistical Inference Problems with V-matrix
6 Solution of Statistical Inference Problems
6.1 Estimation of Conditional Probability Function
6.2 Estimation of Regression Function
6.3 Estimation of Density Ratio Function
6.4 Two-Stage Method for Function Estimation: Data Smoothing and Data Interpolation
6.5 Applications of Density Ratio Estimation
7 Concluding Remarks
References
Statistical Learning and Its Applications
Feature Mapping Through Maximization
of the Atomic Interclass Distances
Adaptive Design of Experiments for Sobol Indices Estimation Based on Quadratic Metamodel
1 Introduction
2 Calculation of Sensitivity Indices Using Quadratic Metamodel
3 Asymptotic Approximation
4 Optimality Criterion and Procedure for Design Construction
5 Experimental Results
6 Conclusion
References
GoldenEye++: A Closer Look into the Black Box
1 Introduction
2 Method
3 Results
4 Concluding Remarks
References
Gaussian Process Regression for Structured Data Sets
1 Introduction
1.1 Approximation Problem
1.2 Factorial Design of Experiments
1.3 Gaussian Process Regression
2 Proposed Approach
3 Experimental Results
4 Conclusion
References
Adaptive Design of Experiments Based on Gaussian Processes
1 Introduction
2 Gaussian Process Regression
3 Adaptive Design of Experiments
4 One-Step Look-Ahead Solutions
4.1 L2 Error Function
4.2 L1 Error Function
4.3 L∞ Error Function
5 Towards More Robust Adaptive DoE Criterion
6 Experiments
6.1 Model Experiments
6.2 Spring Design
7 Conclusions
References
Forests of Randomized Shapelet Trees
Aggregation of Adaptive Forecasting Algorithms Under Asymmetric Loss Function
Visualization and Analysis of Multiple Time Series by Beanplot PCA
1 The Statistical Problem
2 Beanplot Time Series
3 Model-Based Beanplots: Parameterization
4 Multiple Beanplot Time Series
5 Application on Real Data
6 Conclusions
References
Recursive SVM Based on TEDA
1 Introduction
2 SVM Model Formulation
3 TEDA Approach Summary
4 The TEDA SVM Statement
5 TEDA Kernel
6 TEDA SVM Incremental Update
7 Suggestions on the Method Implementation
8 Demonstrations for the Experimental Data
9 Conclusion
References
RDE with Forgetting: An Approximate Solution for Large Values of k with an Application to Fault Detection Problems
Sit-to-Stand Movement Recognition Using Kinect
1 Introduction
2 Movement Recognition: Analysis and Models
3 Experimental Set-Up and Framework
3.1 Data Visualization
3.2 Analysis
4 Evaluation
5 Discussion and Conclusions
References
Additive Regularization of Topic Models for Topic Selection and Sparse Factorization
1 Introduction
2 Additive Regularization of Topic Models
3 Number of Topics Determination
4 Topic Selection in a Sparse Decorrelated Model
5 Conclusions
References
Social Web-Based Anxiety Index's Predictive Information on S&P 500 Revisited
1 Introduction
2 Discussion on the Web Blog Based Anxiety Index
3 Anxiety Index's Predictive Information on the Stock Market, Revisited
4 Conclusion
References
Exploring the Link Between Gene Expression and Protein Binding by Integrating mRNA Microarray and ChIP-Seq Data
1 Introduction
2 Data and Methods
2.1 Description of the Data
2.2 Analysis of ChIP-seq Data
2.3 A Brief Description of MRF Model
2.4 Analysis of Microarray Data
2.5 TSS Selection
3 Results and Discussion
4 Conclusion
References
Evolving Smart URL Filter
in a Zone-Based Policy Firewall for Detecting
Algorithmically Generated Malicious Domains
Lattice-Theoretic Approach to Version Spaces in Qualitative Decision Making
1 Motivation
2 When the Local Utility Functions are Identity Maps
3 When the Local Utility Functions are Known A Priori
4 Concluding Remarks and Further Directions
References
Conformal Prediction and its Applications
A Comparison of Three Implementations of Multi-Label Conformal Prediction
Modifications to p-Values of Conformal Predictors
Cross-Conformal Prediction with Ridge Regression
1 Introduction
2 Conformal and Inductive Conformal Prediction
3 Cross-Conformal Prediction for Regression
4 Normalized Nonconformity Measures
5 Experiments and Results
6 Conclusion
References
Handling Small Calibration Sets in Mondrian Inductive Conformal Regressors
1 Introduction
2 Background
3 Method
4 Results
5 Concluding Remarks
References
Conformal Anomaly Detection of Trajectories with a Multi-class Hierarchy
1 Introduction
2 Method
3 Experiments and Data
3.1 Experiment 1: Comparing pglobal, plocal and ptype Directly
3.2 Experiment 2: Maintaining Computational Cost
3.3 Experiment 3: Wrong Behaviour Type Anomalies
3.4 Experiment 4: Hybrid Rule
4 Results
5 Conclusion
References
Model Selection Using Efficiency of Conformal Predictors
Confidence Sets for Classification
Conformal Clustering and Its Application to Botnet Traffic
Interpretation of Conformal Prediction Classification Models
1 Introduction
2 Method
3 Results
4 Discussion
References
New Frontiers in Data Analysisfor Nuclear Fusion
Confinement Regime Identification
Using Artificial Intelligence Methods
How to Handle Error Bars in Symbolic Regression
for Data Mining in Scientific Applications
1 Introduction
2 The Basic Version of Symbolic Regression via Genetic
Programming
3 Geodesic Distance to Include the Effects of the Error Bars
and Application to Scaling Laws
4 Example of Application: Scaling Laws
4.1 Scaling Laws
4.2 Numerical Results
5 Discussion and Conclusions
References
Applying Forecasting to Fusion Databases
Computationally Efficient Five-Class Image
Classifier Based on Venn Predictors
1 Introduction
2 Venn Predictors
3 Image Pre-processing
4 Results
5 Discussion
References
SOM and Feature Weights Based Method for Dimensionality Reduction in Large Gauss Linear Models
Geometric Data Analysis
Assigning Objects to Classes of a Euclidean Ascending Hierarchical Clustering
1 Assignment by Dichotomies to a System of Classes
2 Distance from Object to Class
3 Assignment Criterion
4 Application to the Survey Data of Trust Barometer (CEVIPOF)
4.1 Data Set
4.2 Structure of the Space of Trust
4.3 Clustering of Individuals
4.4 Assignment of Individuals of Waves 4 and 5
5 Conclusion
References
The Structure of Argument: Semantic Mapping of US Supreme Court Cases
Supporting Data Analytics for Smart Cities: An Overview of Data Models and Topology
1 Introduction
2 Data Models for Smart Cities
3 Topological Queries in Smart Cities
4 Topological Consistency
5 Conclusions
References
Manifold Learning in Regression Tasks
1 Introduction
2 Manifold Learning
2.1 Conventional Manifold Learning Setting
2.2 Manifold Learning as Manifold Reconstruction Problem
2.3 Manifold Learning as Tangent Bundle Manifold Learning
3 Grassmann & Stiefel Eigenmaps
3.1 Structure of the GSE
3.2 GSE: Preliminaries
3.3 GSE: Tangent Manifold Learning Step
3.4 Manifold Embedding Step
3.5 Tangent Bundle Reconstruction Step
4 Solution of the Regression Task
5 Results of Numerical Experiments
References
Random Projection Towards the Baire Metric for High Dimensional Clustering
Optimal Coding for Discrete Random Vector
Author Index