Machine Learning, Optimization, and Data Science 5th International Conference, LOD 2019, Siena, Italy, September 10-13, 2019, Proceedings

• Main Authors: Nicosia, Giuseppe, Pardalos, Panos, Umeton, Renato, Giuffrida, Giovanni, Sciacca, Vincenzo
• Format: eBook
• Language: English
• Published: Netherlands Springer Nature 2020; Springer International Publishing AG; Springer
• Edition: 1
• Series: Lecture Notes in Computer Science
• Subjects: Application software-Testing; Data processing Computer science
• ISBN: 9783030375997; 3030375994; 3030375986; 9783030375980

This book constitutes the post-conference proceedings of the 5th International Conference on Machine Learning, Optimization, and Data Science, LOD 2019, held in Siena, Italy, in September 2019. The 54 full papers presented were carefully reviewed and selected from 158 submissions. The papers cover t...

Table of Contents:

• Intro -- Preface -- Organization -- Contents -- Deep Neural Network Ensembles -- 1 Introduction -- 2 Related Work -- 3 Algorithm -- 3.1 Training -- 3.2 Testing -- 3.3 Larger Models -- 4 Results -- 5 Analysis -- 5.1 ``Good'' Paths -- 5.2 Bounds on Validation Error of Ensembles -- 6 Discussion -- 6.1 Oversampling -- 6.2 Partitions -- 6.3 Justification for Parameters Filtering Clusters -- 7 Conclusion -- References -- Driver Distraction Detection Using Deep Neural Network -- Abstract -- 1 Introduction -- 2 Deep Learning vs Neural Network -- 3 Experiment -- 3.1 Volunteers and Equipment -- 3.2 Data -- 4 Deep Neural Network Results -- 5 Neural Network Results -- 6 Conclusion -- References -- Deep Learning Algorithms for Complex Pattern Recognition in Ultrasonic Sensors Arrays -- 1 Introduction -- 2 Case Study -- 2.1 Dataset Acquisition Bench -- 3 Dataset Visualization -- 4 Preliminary Analysis -- 5 Proposed Methodologies -- 5.1 Independent Time Correlations Analysis -- 5.2 Time Correlations Analysis -- 5.3 Future Works -- 6 Conclusions -- References -- An Information Analysis Approach into Feature Understanding of Convolutional Deep Neural Networks -- 1 Introduction -- 2 Method -- 2.1 Feature Analysis -- 2.2 Salient Information Visualization -- 3 Results and Discussion -- References -- Stochastic Weight Matrix-Based Regularization Methods for Deep Neural Networks -- 1 Introduction -- 2 Motivation -- 3 Proposed Methods -- 3.1 Weight Reinitialization (WMM-WR) -- 3.2 Weight Shuffling (WMM-WS) -- 4 Experimental Setup -- 5 Results and Discussion -- 6 Conclusions -- References -- Quantitative and Ontology-Based Comparison of Explanations for Image Classification -- 1 Introduction -- 2 Deep Learning for Computer Vision -- 3 Towards eXplainable Artificial Intelligence -- 3.1 The Black Box Problem -- 3.2 Explanations -- 3.3 XAI Algorithms
• 1 Introduction
• 7.1 Performance Evaluation for Different Iterations of LIA -- 7.2 Performance Evaluation for the First Iteration of LIA -- 8 Conclusions and Future Work -- References -- Relationship Estimation Metrics for Binary SoC Data -- 1 Introduction -- 2 Previous Work -- 3 Metrics -- 4 Experimental Procedure -- 5 Results and Discussion -- 6 Conclusion -- References -- Network Alignment Using Graphlet Signature and High Order Proximity -- 1 Introduction -- 2 Related Works -- 3 Problem Definition -- 4 GraphletAlign Based Network Alignment -- 4.1 Obtaining Base Mapping -- 4.2 Higher Order Calculation -- 4.3 GraphletAlign Framework -- 5 Experiments and Results -- 5.1 Results and Discussion -- 5.2 Importance of the Order Cell -- 6 Conclusion -- References -- Effect of Market Spread Over Reinforcement Learning Based Market Maker -- Abstract -- 1 Introduction -- 2 Background and Related Work -- 2.1 Financial Markets -- 2.2 Market Making -- 3 Problem Modelling -- 3.1 Environment -- 3.1.1 States -- 3.2 Learning -- 3.3 Reward Function -- 4 Experimental Results -- 4.1 Training Data -- 4.2 Results and Discussion -- 5 Conclusion -- References -- A Beam Search for the Longest Common Subsequence Problem Guided by a Novel Approximate Expected Length Calculation -- 1 Introduction -- 2 State Graph -- 3 Beam Search Framework -- 3.1 Functions for Evaluating Nodes -- 3.2 A Heuristic Estimation of the Expected Length of an LCS -- 3.3 Expressing Existing Approaches in Terms of Our Framework -- 4 Experimental Evaluation -- 5 Conclusions and Future Work -- References -- An Adaptive Parameter Free Particle Swarm Optimization Algorithm for the Permutation Flowshop Scheduling Problem -- 1 Introduction -- 2 Hybrid Adaptive Particle Swarm Optimization -- 3 Results and Discussion -- 4 Conclusions -- References
• 1 Introduction -- 2 Spatial Rule Mining -- 3 Methods and Materials -- 3.1 Geoprocessing -- 3.2 Data Mining -- 4 Results and Discussion -- 5 Conclusions -- References -- On Probabilistic k-Richness of the k-Means Algorithms -- 1 Introduction -- 2 Background and Preliminaries -- 3 k-Richness Properties of k-Means Variants -- 4 k-Means-Random Not Probabilistically Rich -- 5 Experiments -- 6 Discussion -- 7 Concluding Remarks -- References -- Using Clustering for Supervised Feature Selection to Detect Relevant Features -- 1 Introduction -- 2 Objectives -- 3 Methods -- 3.1 K-Medoids Clustering and Determining the Number of Clusters -- 3.2 COLD Algorithm -- 4 Data and Training Procedure -- 4.1 Artificial Examples -- 4.2 Training Procedure -- 5 Results -- 6 Conclusion -- References -- A Structural Theorem for Center-Based Clustering in High-Dimensional Euclidean Space -- 1 Introduction -- 2 Constructing Universal Approximate Centers -- 3 Applications to Center-Based Clustering -- 3.1 k-Clustering -- 3.2 Partial Clustering -- 3.3 Examples -- 3.4 Reduction to Discrete Euclidean Problems -- 4 Conclusion -- References -- Modification of the k-MXT Algorithm and Its Application to the Geotagged Data Clustering -- 1 Introduction -- 2 Clustering Algorithms -- 2.1 k-MXT Algorithm -- 2.2 k-MXT-W Algorithm -- 3 Comparison of Algorithms on Simulated Data -- 3.1 Methodology -- 3.2 The Blobs Type Data -- 3.3 The Circles Data -- 3.4 The Moons Data Type -- 4 Comparison of the Results of the k-MXT and k-MXT-W Algorithms on Real Data Sets -- 5 Conclusions -- References -- CoPASample: A Heuristics Based Covariance Preserving Data Augmentation -- 1 Introduction -- 2 CoPASample Framework -- 2.1 Sample Mean Preservation -- 2.2 Total Covariance Preservation -- 3 Experimental Results -- 4 Conclusion -- References -- Active Matrix Completion for Algorithm Selection
• 4 Heatmap-Based Comparison -- 4.1 Visual Comparison -- 4.2 Metrics -- 4.3 Quantitative Comparison -- 5 Linking Explanations with Structured Knowledge -- 6 Discussion -- 7 Future Work -- References -- About Generative Aspects of Variational Autoencoders -- 1 Introduction -- 2 A Gaussian Mixture Model -- 3 The Variance Law -- 3.1 Experimental Validation -- 4 Moments of the GMM Distribution -- 5 Conclusions -- References -- Adapted Random Survival Forest for Histograms to Analyze NOx Sensor Failure in Heavy Trucks -- 1 Introduction -- 2 Background -- 2.1 Random Survival Forest -- 2.2 Random Forest Classifier for Histogram Data -- 3 Methods -- 3.1 Using Multiple Bins Together for Evaluating Node Split -- 3.2 Weighted Probability of Getting Selected for Split Evaluation Assigned to a Bin -- 4 Empirical Evaluation -- 4.1 Data Preparation -- 4.2 Experiment Setup and Results -- 5 Conclusion and Future Work -- References -- Incoherent Submatrix Selection via Approximate Independence Sets in Scalar Product Graphs -- 1 Introduction -- 2 Incoherent Submatrix Extraction as an Approximate Independent Set Computation -- 3 Relaxing on the Sphere: A New Extraction Approach -- 3.1 The Spectral Estimator -- 3.2 Theoretical Guarantees -- 4 Conclusion and Future Works -- A Minimizing Quadratic Functionals on the Sphere -- A.1 A Semi-explicit Solution -- A.2 Bounds on -- A.3 Perturbation of the Linear Term -- A.4 Neuberger's Theorem -- References -- LIA: A Label-Independent Algorithm for Feature Selection for Supervised Learning -- 1 Introduction -- 2 Literature Review -- 3 Formalizing the Problem -- 4 The Label-Independent Algorithm (LIA) -- 4.1 Step 1: Network-Based Partition -- 4.2 Step 2: Initialization -- 4.3 Step 3: Greedy Selection -- 4.4 Time Complexity Analysis -- 5 Symmetric Uncertainty -- 6 Empirical Study -- 7 Results
• The Measure of Regular Relations Recognition Applied to the Supervised Classification Task -- Abstract -- 1 Human and Machine Cognition -- 2 Probability Measure of Regular Relations Recognition -- 2.1 Related Works -- 2.2 Classification Task -- 3 Experimental Testing of the Measure of Regularities Recognition -- 3.1 Direct Classifier Algorithm Characteristics and Optimisation -- 3.1.1 Working Time -- 3.1.2 Model Size -- 3.2 Back Index ZClassifier Algorithm Complexity -- 4 The Results of the Experiments -- 5 Conclusions -- References -- Simple and Accurate Classification Method Based on Class Association Rules Performs Well on Well-Known Datasets -- Abstract -- 1 Introduction -- 2 Related Work -- 3 Preliminary Concepts -- 4 Problem Statement -- 5 Generating the Complete Set of CARs -- 5.1 Finding the Frequent Itemsets Using Candidate Generation -- 5.2 Generating the Class Association Rules from Frequent Itemsets -- 6 Definition of Our Proposed Method -- 7 Experimental Evaluation -- 8 Conclusion and Future Work -- Acknowledgement -- References -- Analyses of Multi-collection Corpora via Compound Topic Modeling -- 1 Introduction -- 2 A Compound Hierarchical Model -- 3 Posterior Sampling -- 3.1 Inference via Markov Chain Monte Carlo Methods -- 3.2 Empirical Evaluation of Samples -- 4 Experimental Analyses -- References -- Text Mining with Constrained Tensor Decomposition -- 1 Introduction -- 2 Problem Formulation -- 3 The Robust Tensor Power Method -- 4 Proposed Method -- 5 Simulations -- 5.1 Data Generation -- 5.2 Performance -- 6 Conclusion -- References -- The Induction Problem: A Machine Learning Vindication Argument -- 1 Introduction -- 2 The Machine Learning Argument -- 2.1 Example -- 3 Discussion -- 4 Conclusion -- References -- Geospatial Dimension in Association Rule Mining: The Case Study of the Amazon Charcoal Tree -- Abstract