Advances in Computational Intelligence 15th International Work-Conference on Artificial Neural Networks, IWANN 2019, Gran Canaria, Spain, June 12-14, 2019, Proceedings, Part I

• Main Authors: Rojas, Ignacio, Joya, Gonzalo, Catala, Andreu
• Format: eBook
• Language: English
• Published: Cham Springer International Publishing AG 2019
• Edition: 1
• Subjects: Artificial intelligence
• ISBN: 3030205207; 9783030205201

Table of Contents:

• Intro -- Preface -- Organization -- Contents - Part I -- Contents - Part II -- Machine Learning in Weather Observation and Forecasting -- A Deeper Look into 'Deep Learning of Aftershock Patterns Following Large Earthquakes': Illustrating First Principles in Neural Network Physical Interpretability -- Abstract -- 1 Introduction -- 2 Artificial Neural Networks in Statistical Seismology -- 2.1 Literature Survey -- 2.2 The DeVries18 Study -- 3 Applying First Principles to Neural Network Interpretability -- 3.1 Was High Abstraction Required to Predict Aftershock Patterns? -- 3.2 Were Stress Metrics the Most Pertinent Physical Parameters? -- 4 Conclusions -- References -- Boosting Wavelet Neural Networks Using Evolutionary Algorithms for Short-Term Wind Speed Time Series Forecasting -- Abstract -- 1 Introduction -- 2 Structure of the Wavelet Neural Network -- 2.1 The Framework of the Network -- 2.2 Ridge Type Wavelet Basis Function -- 2.3 Training of the Network -- 3 Network Training -- 3.1 Evolutionary Algorithms -- 3.2 Coordinate Dictionary Search Optimization (CDSO) Algorithm -- 4 Case Study - Wind Speed Forecasting -- 4.1 The Model -- 4.2 Model Performance -- 5 Conclusion -- Acknowledgments -- References -- An Approach to Rain Detection Using Sobel Image Pre-processing and Convolutional Neuronal Networks -- Abstract -- 1 Introduction -- 2 Methods -- 2.1 Dataset -- 2.2 Preprocessing -- 3 Experiments and Results -- 4 Discussion and Future Work -- Acknowledgements -- References -- On the Application of a Recurrent Neural Network for Rainfall Quantification Based on the Received Signal from Microwave Links -- Abstract -- 1 Introduction -- 2 Experimental Setup and Data Processing -- 3 Feature Extraction -- 3.1 Variability of the RSL with Time -- 3.2 Feature Extraction Module -- 4 Rainfall Quantification Using a Recurrent Neural Network
• A Scalable Long-Horizon Forecasting of Building Electricity Consumption -- 1 Introduction -- 2 Data Pre-processing -- 2.1 Missing Values and Outliers in Power Data -- 2.2 Data Synchronization -- 3 Hybridized Recursive-Direct (HRD) Multi-step Ahead Forecast -- 3.1 Direct Forecast -- 3.2 Recursive Forecast -- 3.3 Hybridized Direct-Recursive (HDR) Forecast -- 4 Buildings' Month Ahead Load Forecasting -- 4.1 Feature Derivation and Selection -- 4.2 HDR Based Month Ahead Forecasting Using ANN and SVR -- 4.3 Results -- 5 Quarter Ahead Load Forecasting -- 5.1 Feature Extraction -- 5.2 HDR Based Quarter Ahead Forecasting Using Linear Regression -- 5.3 Results -- 6 Discussion -- 7 Conclusion -- References -- Long-Term Forecasting of Heterogenous Variables with Automatic Algorithm Selection -- 1 Introduction -- 2 Data Pre-processing -- 2.1 Granger's Causality Test -- 2.2 Data Transformation and Predictor Matrix -- 3 Time Series Modelling Using Machine Learning Algorithms -- 3.1 Artificial Neural Networks -- 3.2 Support Vector Regression -- 3.3 Random Forests -- 4 Real-Time Switch for Automatic Algorithm Selection -- 4.1 Real-Time Switch -- 5 Results and Discussion -- 6 Conclusion -- References -- Automatic Time Series Forecasting with GRNN: A Comparison with Other Models -- 1 Introduction -- 2 Generalized Regression Neural Networks -- 3 Time Series Forecasting with GRNN -- 3.1 Preprocessing -- 3.2 Autoregressive Lags and Number of Neurons -- 3.3 Selecting the Smoothing Parameter -- 3.4 Multi-step Ahead Strategy -- 4 Automatic Time Series Forecasting in R -- 4.1 Computational Intelligence Methods -- 4.2 Statistical Models -- 4.3 A Combination of Methods -- 5 Experimentation -- 6 Conclusions -- References -- Improving Online Handwriting Text/Non-text Classification Accuracy Under Condition of Stroke Context Absence -- 1 Introduction -- 2 Proposed Solution
• 4 Experiments and Results
• 5 Experiments and Results -- 6 Conclusions and Future Works -- Acknowledgments -- References -- Ambient Temperature Estimation Using WSN Links and Gaussian Process Regression -- 1 Introduction -- 2 Related Work -- 3 Methodology -- 3.1 Gaussian Processes for Machine Learning -- 3.2 Experimental Setup -- 4 Experimental Results -- 5 Conclusions -- References -- Computational Intelligence Methods for Time Series -- Voice Command Recognition Using Statistical Signal Processing and SVM -- Abstract -- 1 Introduction -- 2 Command Recognition System -- 2.1 Data Base of Commands -- 2.2 Signal Preprocessing for Feature Extraction -- 2.3 Support Vector Machine Classifier -- 2.4 Numerical Results of Command Recognition -- 3 Speaker Identification Using Developed System -- 3.1 Data Base -- 3.2 Details of System -- 3.3 Results of Experiments -- 4 Conclusions -- References -- Enterprise System Response Time Prediction Using Non-stationary Function Approximations -- 1 Introduction -- 2 Problem Description -- 3 Model Development -- 3.1 Data Preprocessing -- 3.2 New Feature Identification -- 3.3 Modeling Input Forecasts -- 3.4 Prediction Models -- 4 Experimental Results and Discussion -- 5 Conclusions -- References -- Using Artificial Neural Networks for Recovering the Value-of-Travel-Time Distribution -- Abstract -- 1 Introduction -- 2 Methodology -- 2.1 Preliminary -- 2.2 Uncovering Individual VTTs Using ANNs -- 2.3 ANN Development -- 3 Application to Real VTT Data -- 3.1 Training and Simulation -- 3.2 Results -- 4 Cross-validation -- 5 Conclusions and Discussion -- References -- Sparse, Interpretable and Transparent Predictive Model Identification for Healthcare Data Analysis -- Abstract -- 1 Introduction -- 2 Model Representation -- 3 Sparse Dictionary Learning and NARMAX Model Estimation -- 4 Case Studies and Real Applications
• 3 Evaluation of Our Solution -- 4 Conclusion -- References -- Improving Classification of Ultra-High Energy Cosmic Rays Using Spacial Locality by Means of a Convolutional DNN -- 1 Introduction and Problem Description -- 2 Data Description -- 2.1 Data Preprocessing -- 3 Methodology: Feed-Forward Neural Network (FFNN) and Convolutional Neural Network (CNN) -- 3.1 FFNN -- 3.2 CNN -- 4 Results -- 5 Conclusions -- References -- Model and Feature Aggregation Based Federated Learning for Multi-sensor Time Series Trend Following -- 1 Introduction -- 2 Multi-sensor TSD Trend Following with Model and Feature Aggregation in Federated Learning -- 2.1 Traditional Federated Learning -- 2.2 Federated Learning Based Model and Feature Fusion for Multi-sensor TSD -- 2.3 Global Aggregated Feature Based Trend Following of TSD -- 3 Experiments and Analysis -- 3.1 Settings for Local Feature Extraction -- 3.2 Settings for Feature Aggregation and Feature Based Trend Following -- 3.3 Experimental Performance -- 4 Conclusions -- References -- Robust Echo State Network for Recursive System Identification -- 1 Introduction -- 2 Fundamentals of the Echo State Network -- 2.1 Recursive Algorithms for Parameter Estimation -- 3 Methodology of Evaluation and Simulation -- 3.1 Evaluation and Simulation -- 4 Results -- 5 Conclusions and Further Work -- References -- Random Hyper-parameter Search-Based Deep Neural Network for Power Consumption Forecasting -- 1 Introduction -- 2 Related Work -- 3 Methodology -- 3.1 Hyper-parameters Tuning -- 3.2 Multi-step to Single-step Regression -- 3.3 Smoothing Filter -- 4 Results -- 4.1 Dataset Description -- 4.2 Error Metrics -- 4.3 Performance in Terms of Error -- 5 Conclusions -- References -- A First Approximation to the Effects of Classical Time Series Preprocessing Methods on LSTM Accuracy -- 1 Introduction -- 2 LSTM -- 3 Preprocessing Methods
• 4.1 The Relation Between Influenza-Like Illness Incidence Rate and Deaths -- 4.2 Analysis of Beijing Air Quality -- 5 Conclusion -- Acknowledgments -- References -- Use of Complex Networks for the Automatic Detection and the Diagnosis of Alzheimer's Disease -- 1 Introduction -- 2 Methods -- 3 Network Measures -- 3.1 Clustering Coefficient -- 3.2 Mean Jump Length -- 3.3 Betweenness Centrality -- 4 Data -- 5 Results -- 6 Conclusions -- References -- The Generalized Sleep Spindles Detector: A Generative Model Approach on Single-Channel EEGs -- 1 Introduction -- 2 A Problem Beyond Detection -- 3 Methods -- 3.1 The Discriminative Embedding Transform -- 3.2 MDL-Based Clustering -- 4 Results -- 5 Conclusion -- References -- DeepTrace: A Generic Framework for Time Series Forecasting -- 1 Introduction -- 2 Preliminary -- 2.1 Autocorrelation -- 2.2 1-D Convolutions -- 2.3 Data Preparation -- 2.4 Metrics -- 3 Model Architecture -- 3.1 Convolutional Block (CB) -- 3.2 Recurrent Block (RB) -- 3.3 Linear Block (LB) -- 3.4 Residual Connections -- 3.5 Model Variants -- 4 Training and Testing Phase -- 5 Optimization -- 6 Experimentation -- 7 Conclusions -- References -- Automatic Identification of Interictal Epileptiform Discharges with the Use of Complex Networks -- 1 Introduction -- 2 Methods -- 3 Network Measures -- 3.1 Strongly Connected Component (SCC) -- 3.2 Average Shortest Path Length (L) -- 3.3 Mean Jump Length () -- 4 Data -- 5 Results -- 6 Conclusions -- References -- Anomaly Detection for Bivariate Signals -- 1 Introduction -- 2 Methodology -- 2.1 Clustering -- 2.2 Introducing Reference Curves for Summarizing the Clusters -- 2.3 Time-Series Realignment Within Clusters -- 2.4 Anomaly Detection -- 3 An Experimental Example on Simulated Data -- 4 An Application to Real-World Data -- 5 Conclusion -- References