Machine Learning, Optimization, and Data Science

Name: Machine Learning, Optimization, and Data Science | 6th International Conference, LOD 2020, Siena, Italy, July 19-23, 2020, Revised Selected Papers, Part I
Brand: Springer
Price: 96.29 EUR
Availability: OnlineOnly

6th International Conference, LOD 2020, Siena, Italy, July 19-23, 2020, Revised Selected Papers, Part I

Giuseppe Nicosia Varun Ojha Emanuele La Malfa Giorgio Jansen Vincenzo Sciacca Panos Pardalos Giovanni Giuffrida Renato Umeton(Editor)

Springer (Publisher)

Published on 7. January 2021

XXXVII, 740 pages

E-Book

PDF with digital watermarking

System requirements

978-3-030-64583-0 (ISBN)

€96.29incl. 7% vat

System requirements

for PDF with digital watermarking

E-Book Single Licence

Available for download

Description

More details

Other editions

Content

Intro
Preface
Organization
Contents - Part I
Contents - Part II
Revisiting Clustering as Matrix Factorisation on the Stiefel Manifold
1 Introduction
1.1 Historical Background
1.2 Our Contribution
1.3 Notation
2 Non-negative Factorisation of the Stiefel Manifold
2.1 Model
2.2 Ideal Solution
2.3 The Latent Variable Model
2.4 Generalized Posterior and Estimator
2.5 A PAC-Bayesian-Flavored Error Bound
3 A Langevin Sampler
3.1 Computing the Gradient on the Stiefel Manifold
3.2 The Alternating Langevin Sampler
4 Numerical Experiment
References
A Generalized Quadratic Loss for SVM and Deep Neural Networks
1 Introduction
2 Related Works
3 The Modified Loss SVM Problem
4 Notation
5 Algorithms
5.1 The SMOS Optimization Algorithm
5.2 The RTS Optimization Algorithm
5.3 The Deep Learning Framework
6 Results
7 Conclusions
References
Machine Learning Application to Family Business Status Classification
1 Introduction
2 Description of the Dataset
3 Data Pre-processing
4 Application of Machine Learning Techniques
5 Conclusions and Possible Future Developments
References
Using Hessians as a Regularization Technique
1 Introduction
2 Proposed Method
3 Experiment
4 Results
5 Conclusion
References
Scaling Up Quasi-newton Algorithms: Communication Efficient Distributed SR1
1 Introduction
2 Sampled Limited-Memory SR1 (S-LSR1)
2.1 Naive Distributed Implementation of S-LSR1
3 Efficient Distributed S-LSR1 (DS-LSR1)
3.1 Reducing the Amount of Information Communicated
3.2 Balancing the Computation Across the Nodes
3.3 The Distributed S-LSR1 (DS-LSR1) Algorithm
3.4 Complexity Analysis - Comparison of Methods
4 Numerical Experiments
4.1 Scaling
4.2 Performance of DS-LSR1
5 Final Remarks
References
Should Simplicity Be Always Preferred to Complexity in Supervised Machine Learning?
1 Introduction
2 Theoretical Analysis
References
An Application of Machine Learning to Study Utilities Expenses in the Brazilian Navy
1 Introduction
2 Related Work
3 Methodology
3.1 Data
3.2 Methods
4 Experimental Study
4.1 Experiments for the Model Selection
4.2 Performance Results
4.3 Feature Importance
4.4 Reduced Dataset Analysis
4.5 Potential Uses
5 Conclusions
References
The Role of Animal Spirit in Monitoring Location Shifts with SVM:Novelties Versus Outliers
1 Motivation
2 The Basic Idea
2.1 Novelty Versus Outlier Detection
3 Assessing Survey Data
3.1 Structural Breaks
4 Detecting Novelties with SVM
4.1 Choosing a Training Sample
4.2 Novelties Since June 2019
5 Rare Event in March-April-May 2020
6 Concluding Remarks
References
Long-Term Prediction of Physical Interactions: A Challenge for Deep Generative Models
1 Introduction
2 Temporal Deep Generative Models
3 Training and Testing Methods
4 Results
5 Conclusion
References
Semantic Segmentation of Neuronal Bodies in Fluorescence Microscopy Using a 2D+3D CNN Training Strategy with Sparsely Annotated Data
1 Introduction and Related Work
2 Methodology
3 Evaluation and Results
4 Conclusions
References
Methods for Hyperparameters Optimization in Learning Approaches: An Overview
1 Introduction
2 Hyperparameter Optimization
3 Overview on Existing Methods
4 Gradient-Based Methods
5 Discussion
References
Novel Reconstruction Errors for Saliency Detection in Hyperspectral Images
1 Introduction
2 Reconstruction Error Measures
2.1 Mathematical Properties
3 Structure of the Algorithm
3.1 NMF
3.2 Clustering Stage
4 Experimental Results
4.1 HS-SOD Dataset
4.2 Hermiston Dataset
5 Conclusions
References
Sparse Consensus Classification for Discovering Novel Biomarkers in Rheumatoid Arthritis
1 Introduction
2 Methods
2.1 Rheumatologic Transcriptomic Data
2.2 Sparse Logistic Regression
2.3 Bayesian Networks
3 Results and Discussion
3.1 Identification of Response Biomarkers
3.2 Identification of Protein-Protein Interactions
4 Conclusions
References
Learning More Expressive Joint Distributions in Multimodal Variational Methods
1 Introduction
2 Related Work
3 Variational Inference
4 Learning Flexible and Complex Distributions in Multimodal Variational Methods
5 Evaluation
5.1 Evaluation of the Model and Generated Sample Quality
5.2 Image Transformation Tasks
6 Conclusion and Future Work
References
Estimating the F1 Score for Learning from Positive and Unlabeled Examples
1 Introduction
2 Problem Formulation
3 Literature Review
3.1 PU Learning Algorithms: Two-Step Strategy
3.2 Performance Estimation
4 Estimating the F1 Score
4.1 Approach to Estimate F1-score
4.2 Estimating
4.3 Behavior Under Noisy
5 Experimental Setup
5.1 Datasets and Setup
5.2 Experimental Conditions and Performance Metrics
6 Results
6.1 Checking the Assumptions
6.2 Correct
6.3 Noisy
7 Conclusion
References
Dynamic Industry-Specific Lexicon Generation for Stock Market Forecast
1 Introduction
2 Related Work
3 Problem Formulation
4 The Proposed Approach
4.1 Generation of Industry-Specific Lexicons
4.2 Class Prediction Algorithm
5 Experimental Framework
5.1 Datasets
5.2 Methodology
5.3 Results
6 Conclusions and Future Work
References
A Stochastic Optimization Model for Frequency Control and Energy Management in a Microgrid
1 Introduction
2 Context of the Problem
2.1 Frequency Containment Reserve - FCR
2.2 Related Work
3 A Stochastic Day-Ahead Optimization Model
4 MPC Controller and Simulation
5 Numerical Assessments
5.1 Sensitivity Analysis and Out-of-Sample Simulations
5.2 Closed-Loop Simulations
6 Conclusion
References
Using the GDELT Dataset to Analyse the Italian Sovereign Bond Market
1 Introduction
2 Related Work
3 Data
3.1 About GDELT
3.2 Yield Spread
4 Methods
4.1 Big Data Management
4.2 Feature Engineering
4.3 Big Data Analytics
4.4 Experimental Analysis
5 Conclusions
References
Adjusted Measures for Feature Selection Stability for Data Sets with Similar Features
1 Introduction
2 Concepts and Methods
2.1 Feature Selection Stability
2.2 Adjusted Stability Measures
3 Experiments and Results
3.1 Experimental Results on Artificial Feature Sets
3.2 Experimental Results on Real Feature Sets
4 Conclusions
References
Reliable Solution of Multidimensional Stochastic Problems Using Metamodels
1 Introduction
2 Concepts and Methods
2.1 Kriging
2.2 Pareto Set and Pareto Frontier
2.3 Attainment Function
3 Estimating Expectation and Standard Deviation with Metamodels
3.1 General Idea
3.2 Estimation of Expectation and Standard Deviation
4 Experiments
4.1 Design of Experiments
4.2 Computational Aspects
5 Evaluation of the Experiments
6 Conclusion
References
Understanding Production Process Productivity in the Glass Container Industry: A Big Data Approach
1 Introduction
2 The Glass Container Production Process
3 Methodology
4 On Going Work
References
Random Forest Parameterization for Earthquake Catalog Generation
1 Introduction
2 Data
3 Methodology
3.1 Data Processing
3.2 Random Forest
4 Experiments
5 Conclusions
References
Convolutional Neural Network and Stochastic Variational Gaussian Process for Heating Load Forecasting
1 Introduction
2 Materials and Methods
2.1 Problem Definition and System Overview
2.2 Dataset
2.3 Convolutional Neural Network Model
2.4 Stochastic Variational Gaussian Process Model
2.5 Performance Measure
3 Results
3.1 CNN Model
3.2 SVGP Model
3.3 Model Comparison
4 Conclusion and Ongoing Work
References
Explainable AI as a Social Microscope: A Case Study on Academic Performance
1 Introduction
2 Data and Pre-processing
2.1 FFC Dataset
2.2 Pre-processing and Feature Selection
3 Comparative Analysis
3.1 General Indicators
3.2 Proposed Methodology: Targeted Indicators
3.3 Results and Discussion
4 Future Work
5 Concluding Remarks
References
Policy Feedback in Deep Reinforcement Learning to Exploit Expert Knowledge
1 Introduction
2 Design
3 Experiments and Results
4 Discussion
5 Conclusions
References
Gradient Bias to Solve the Generalization Limit of Genetic Algorithms Through Hybridization with Reinforcement Learning
1 Introduction
2 Related Work
3 The Generalization Limit of Genetic Algorithms
3.1 Experiments
3.2 Discussion
4 X-DDPG
4.1 Motivations
4.2 The Algorithm
4.3 Experiments and Results
5 Conclusions
5.1 Limits and Future Directions
References
Relational Bayesian Model Averaging for Sentiment Analysis in Social Networks
1 Introduction
2 Literature Review
3 Background Concept and Model Training
4 RBMA: The Predictive Model
5 Computational Results
6 Conclusions
References
Variance Loss in Variational Autoencoders
1 Introduction
2 Variational Autoencoders
2.1 KL Divergence in Closed Form
3 The Variance Loss Issue
3.1 General Case
4 Addressing the Variance Loss
5 Conclusions
References
Wasserstein Embeddings for Nonnegative Matrix Factorization
1 Introduction
2 Optimal Transport and Wasserstein Distance
3 Cuturi Regularized Optimal Transport (Discrete)
4 Wasserstein Embeddings NMF (WE-NMF)
5 Experiments
5.1 Settings
5.2 Other Optimal Transport Algorithms
5.3 Empirical Results
6 Conclusion
References
Investigating the Compositional Structure of Deep Neural Networks
1 Introduction
2 Related Work
3 Methods
3.1 Basic Definitions
3.2 From activation patterns to the APD
3.3 Clustering the input dataset using the APD
4 Results
5 Conclusions and Future Developments
References
Optimal Scenario-Tree Selection for Multistage Stochastic Programming
1 Introduction
2 Related Works
3 Methodology
3.1 Multistage Stochastic Programming
3.2 Inference of a Decision Policy from a Scenario-Tree Solution
3.3 Quality Evaluation of a Decision Policy
4 Proposed Approach
5 Numerical Tests
5.1 The Benchmark Problem
5.2 Computational Results
6 Conclusions
References
Deep 3D Convolution Neural Network for Alzheimer's Detection
1 Introduction
2 Related Work
3 Proposed Work
3.1 Dataset and Pre-processing
3.2 Model
3.3 Training
4 Experiments
5 Conclusion and Future Work
References
Combinatorial Reliability-Based Optimization of Nonlinear Finite Element Model Using an Artificial Neural Network-Based Approximation
1 Introduction
2 Optimized Structural Member
3 Optimization Task
3.1 Combinatorial Optimization Problem - Design Space Definition
3.2 Stochastic Model
3.3 Sensitivity Analysis
3.4 The ANN Surrogate Models
3.5 Optimization Problem, Decision-Making Criteria and Results
4 Conclusions
References
CMAC: Clustering Class Association Rules to Form a Compact and Meaningful Associative Classifier
1 Introduction and Scientific Background
2 Related Work
3 Problem Definition and Goals
4 Our Proposed Method - CMAC
4.1 Finding "Strong" Class Association Rules
4.2 The New "Direct" Distance Metric
4.3 Clustering
4.4 Extracting the "Representative" CARs
5 Experimental Setting and Results
6 Conclusion and Future Work
References
GPU Accelerated Data Preparation for Limit Order Book Modeling
1 Introduction
1.1 The Limit Order Book
2 Previous Works
3 Proposed Methods
3.1 Data Collection
3.2 Data Preparation
4 Evaluation and Results
4.1 Used Data
4.2 Hardware and Software Architecture
4.3 Execution Time
4.4 Inference
5 Application and Conclusions
References
Can Big Data Help to Predict Conditional Stock Market Volatility? An Application to Brexit
1 Introduction
2 Data and Methodology
2.1 The News Related Variables
2.2 The News Augmented GARCH Models
3 Empirical Results
3.1 In-Sample Forecasting Results
3.2 Out-of-Sample Forecasting Results
4 Conclusion
References
Importance Weighting of Diagnostic Trouble Codes for Anomaly Detection
1 Introduction
2 Analysis of a Sample Vehicle
3 Anomaly Detection with VAEs
3.1 Formal Background
3.2 Implementation
3.3 Anomaly Detection
4 Importance Weighting of DTCs
5 Related Work
6 Conclusion
References
Identifying Key miRNA-mRNA Regulatory Modules in Cancer Using Sparse Multivariate Factor Regression
1 Introduction
1.1 Literature Review
1.2 Our Contributions
2 Problem Definition
3 Method
3.1 Sparse Multivariate Factor Regression
4 Materials
5 Results and Discussion
5.1 Experimental Settings
5.2 Predictive Performance Comparison
5.3 Extracting Key miRNA-mRNA Regulatory Modules
5.4 Literature Validation of Identified miRNA-mRNA Regulatory Modules
6 Conclusions
References
A Krill Herd Algorithm for the Multiobjective Energy Reduction Multi-Depot Vehicle Routing Problem
1 Introduction
2 Parallel Multi-Start Non-dominated Sorting Krill Herd Algorithm (PMS-KH)
3 Computational Results
4 Conclusions and Future Research
References
Optimal Broadcast Strategy in Homogeneous Point-to-Point Networks
1 Motivation
2 Problem
3 Solution
4 Discussion
5 Conclusions and Trends for Future Work
References
Structural and Functional Representativity of GANs for Data Generation in Sequential Decision Making
1 Introduction
2 Related Work
3 Method
3.1 Preliminaries
3.2 Reinforcement Learning
3.3 Generating Experiences Using GANs
3.4 Evaluation Framework
4 Experimental Setup
4.1 Specific Research Questions
4.2 Evaluation Environments
4.3 Behaviour Policies and Datasets
4.4 GAN Architecture and Hyperparameters
5 Results
5.1 Generating One-Step Horizon Sequential Data
5.2 Structural Representativity
5.3 Functional Representativity
6 Conclusion
References
Driving Subscriptions Through User Behavior Modeling and Prediction at Bloomberg Media
1 Introduction
2 Subscription Propensity Model
3 Elastic Paywall Optimization
4 Results
Automatic Setting of DNN Hyper-Parameters by Mixing Bayesian Optimization and Tuning Rules
1 Introduction
2 Bayesian Optimization
3 Network Behaviour Analysis
3.1 Execution Pipeline
3.2 Diagnosis Module
3.3 Tuning Rules
4 Experiments
5 Conclusion and Future Work
References
A General Approach for Risk Controlled Trading Based on Machine Learning and Statistical Arbitrage
1 Introduction
2 Related Work
3 Problem Formulation
4 The Proposed Approach
5 Feature Engineering
6 Baselines
7 Ensemble
8 Dynamic Asset Selection
9 Experimental Framework
10 Conclusions and Future Work
References
Privacy Preserving Deep Learning Framework in Fog Computing
1 Introduction
2 Related Work
3 Framework
4 Performance Evaluation
4.1 Dataset
4.2 Description Models
4.3 Experiments
4.4 Results
5 Conclusion
References
On the Reliability of Dynamical Stochastic Binary Systems
1 Introduction
2 Dynamical Stochastic Binary Systems
3 Interplay Between SBS and DSBS
4 Reliability Evaluation
5 Proof-of-Concept
6 Discussion
7 Concluding Remarks
References
A Fast Genetic Algorithm for the Max Cut-Clique Problem
1 Motivation
2 Computational Complexity
3 GRASP/VND Heuristic
4 Genetic Algorithm
4.1 Fitness and Notation
4.2 Selection: Tournament Selection
4.3 Crossover
4.4 Mutation
4.5 Replacement: New Generation
4.6 Feasible Solution
4.7 Parameters Adjustment
5 Computational Results
6 Conclusions and Trends for Future Work
References
The Intellectual Structure of Business Analytics: 2002-2019
1 Introduction
2 Literature Review
3 Analyzing the Research
3.1 Analysis Methods and Data Collection
3.2 Analyzing with CiteSpace
4 Results of the Analysis
4.1 Clustering Analysis
4.2 Key Documents
4.3 Burst Detection
4.4 Timeline Analysis
5 Conclusion
References
A New Approach to Early Warning Systems for Small European Banks
1 Introduction
2 Literature Review
3 Data and Sample of Distress Events
4 Methodology
4.1 Data Pre-processing
4.2 Decision Trees
5 Results
6 Conclusions
References
Evaluating the Impact of Training Loss on MR to Synthetic CT Conversion
1 Introduction
2 Related Work
3 Materials and Methods
3.1 Dataset Description
3.2 Proposed Approach
3.3 Losses Description
4 Experiments
4.1 Benchmark Platform
4.2 Deep Convolutional Neural Network (DCNN) Model for MR to sCT
4.3 Training
4.4 Implementation of Losses for MR-to-sCT Conversion
4.5 Metric for Result Evaluation
5 Results
6 Discussion
7 Conclusion
References
Global Convergence of Sobolev Training for Overparameterized Neural Networks
1 Introduction
2 Main Result
2.1 Consequences for a Network with Bias
2.2 Discussion
3 Proof of Theorem 1
3.1 Analysis Near Initialization
3.2 Proof of Global Convergence
A Supplementary proofs for Sect.3.1
B Proof of Proposition 1
References
Injective Domain Knowledge in Neural Networks for Transprecision Computing
1 Introduction
2 Related Works
3 Domain Knowledge Injection
4 Transprecision Computing
4.1 Problem Description
4.2 Data Set Creation
4.3 Knowledge Injection
5 Experimental Evaluation
5.1 Models Accuracy
5.2 Semantic Based Regularization Impact
6 Conclusion and Future Works
References
XM_HeatForecast: Heating Load Forecasting in Smart District Heating Networks
1 Introduction
2 XM_HeatForecast: System Overview
2.1 Main Modules and Software Structure
2.2 Testing Dataset
2.3 Data Pre-processing
3 Model Generation and Forecasting
3.1 Model Generator
3.2 Forecaster
3.3 Performance Evaluation
4 Model and Forecast Visualization and Analysis
4.1 Visualization of Heating Load Predictions
4.2 Visualization of Model and Prediction Performance
4.3 Visualization of Model Parameters
4.4 Visualization and Analysis of Training Data
5 Communication Interface
5.1 Back-End
5.2 Front-End
6 Case Study
7 Conclusions
References
Robust Generative Restricted Kernel Machines Using Weighted Conjugate Feature Duality
1 Introduction
2 Weighted Restricted Kernel Machines
2.1 Weighted Conjugate Feature Duality
2.2 Generation
3 Robust Estimation of the Latent Variables
3.1 Robust Weighting Scheme
3.2 Algorithm
4 Experiments
5 Conclusion
References
A Location-Routing Based Solution Approach for Reorganizing Postal Collection Operations in Rural Areas
1 Introduction
2 Problem Description
3 Problem Formulation
4 An IP Based Solution Method
4.1 Lower Bound Determination
4.2 Upper Bound Determination
5 Computational Results
6 Conclusions
References
Dynamic Selection of Classifiers Applied to High-Dimensional Small-Instance Data Sets: Problems and Challenges
1 Introduction
2 Background
2.1 Dynamic Selection
2.2 High-Dimensional Data
2.3 Feature Selection
3 Experiments
3.1 Data Sets
3.2 Experimental Design
3.3 Comparison of Techniques
4 Results and Discussion
5 Conclusions
References
Caching Suggestions Using Reinforcement Learning
1 Introduction
2 Data Lake Architecture
3 Related Works
4 Proposed Approach
4.1 Evaluation Measures
5 Use Case
5.1 Data Analysis
6 Experimental Results
7 Conclusions
References
Image Features Anonymization for Privacy Aware Machine Learning
1 Introduction
2 Methodology
2.1 Data Description
2.2 Feature Extraction Techniques
2.3 Feature Anonymization Techniques: Differential Privacy
2.4 Utility Evaluation : Machine Learning Techniques
2.5 Privacy Validation
3 Evaluation and Analysis
3.1 Image Anonymization of MNIST and CIFAR-10 Data
3.2 Feature Anonymization of Dogs&Cats Data
3.3 Feature Anonymization of Wikiset and LFW Data
4 Conclusions and Future Work
References
Prediction of Spot Prices in Nord Pool's Day-Ahead Market Using Machine Learning and Deep Learning
1 Introduction
2 Related Works
3 Dataset
3.1 Data Sources and Feature Engineering and Importance
4 Predicting the Day-Ahead Hourly Prices
4.1 Autoregression
4.2 Prophet
4.3 Onestep LSTM
4.4 Encoder-Decoder LSTM
4.5 CNN-LSTM
4.6 Ensemble Learning
4.7 Forecast Accuracy Evaluation
5 Results
6 Conclusion
References
Unit Propagation by Means of Coordinate-Wise Minimization
1 Introduction
1.1 Max-Sum Problem and Arc Consistency
2 LP Relaxation of Weighted Partial Max-SAT
2.1 Coordinate-Wise Minimization
2.2 Feasibility of Primal Solutions
3 LP Relaxation of SAT
3.1 Unit Propagation
4 Propagation in Weighted Max-SAT
5 Conclusion
A Proof of Theorem 2
References
A Learning-Based Mathematical Programming Formulation for the Automatic Configuration of Optimization Solvers
1 Introduction
2 The Algorithm Configuration Problem
3 The PMLP and the CSSP
3.1 Performance Map Learning Phase
3.2 Configuration Space Search Problem
4 Experimental Results
4.1 Building the Dataset
4.2 PMLP Experimental Setup
4.3 CSSP Experimental Setup
4.4 Results
5 Conclusions
References
Reinforcement Learning for Playing WrapSlide
1 Introduction
2 Methodological Approach
2.1 Symmetries, Equivalence Classes and Tilings on a Torus
2.2 Reinforcement Learning
2.3 A Mathematical Implementation of the WrapSlide Puzzle
2.4 The Reinforcement Learning Agent Configuration
3 Numerical Results
3.1 The Reinforcement Learning Training Procedure
3.2 The Reinforcement Learning Testing Procedure
4 Conclusion
References
Discovering Travelers' Purchasing Behavior from Public Transport Data
1 Introduction
2 Related Work
3 Problem Definition
4 Proposed Methodology
4.1 Steps 1-2: Web Scraping and Process Mining
4.2 Step 3: Discovery of Purchase Factors
4.3 Step 4: Prediction Model
5 A Case Study
5.1 Steps 1-2: Web Scraping and Process Mining
5.2 Step 3: Discovery of Purchase Factors
5.3 Step 4: Prediction Model
6 Conclusions
References
Correction to: Machine Learning, Optimization, and Data Science
Correction to: G. Nicosia et al. (Eds.): Machine Learning, Optimization, and Data Science, LNCS 12565, https://doi.org/10.1007/978-3-030-64583-0
Correction to: Machine Learning, Optimization, and Data Science
Correction to: G. Nicosia et al. (Eds.): Machine Learning, Optimization, and Data Science, LNCS 12565, https://doi.org/10.1007/978-3-030-64583-0
Author Index

System requirements

Save as PDF Copy link into clipboard

Schweitzer Fachinformationen

Machine Learning, Optimization, and Data Science

Description

More details

Other editions

Additional editions

Content

System requirements