Theory and Applications of Time Series Analysis: Selected Contributions from ITISE 2022 (Contributions to Statistics)

This book presents the latest developments in the theory and applications of time series analysis and forecasting. Comprising a selection of refereed papers, it is divided into several parts that address modern theoretical aspects of time series analysis, forecasting and prediction, with applications to various disciplines, including econometrics and energy research. The broad range of topics discussed, including matters of particular relevance for sustainable development, will give readers a modern perspective on the subject. The included contributions were originally presented at the 8th International Conference on Time Series and Forecasting, ITISE 2022, held in Gran Canaria, Spain, June 27-30, 2022. The ITISE conference series provides a forum for scientists, engineers, educators and students to discuss the latest advances and implementations in the foundations, theory, models and applications of time series analysis and forecasting. It focuses on interdisciplinary research encompassing computer science, mathematics, statistics and econometrics. Preface Contents Theoretical Aspects of Time Series Online Estimation Methods for Irregular Autoregressive Models 1 Introduction 2 Methods 2.1 Irregular Autoregressive Model 2.2 Online Estimation Algorithms 3 Simulation Experiments 4 Application to Real-Life Data 4.1 Flow of the River Nile 4.2 Infant Heart Rate 4.3 Brightness of Astronomical Object 5 Discussion References Costationary Whitenoise Processes and Local Stationarity Testing 1 Introduction 2 Costationary Whitenoise Processes 3 Estimation of CDF for Innovations 4 A Non-parametric Bootstrap Stationarity Test 5 Simulation Studies 6 Stationarity Testing of Equity Returns References Econometrics The Effects of Healthy and Sustainable Transportation, Commerce, and Spillover on Airbnb Performance 1 Introduction 2 Literature Review 3 Materials 3.1 Study Area 3.2 Data and Variables 4 Methods 5 Findings 6 Discussion and Conclusion References A Semi-Markov Approach to Financial Modelling During the COVID-19 Pandemic 1 Introduction 2 The Model 3 Application to Financial Data 4 Conclusion References From News to Sentiments and Stock Price Directions 1 Introduction 2 What Contribute to Sentiment and Share Prices? 3 Analysing Stock Price Trends 4 A Case Study on How News Sentiment Affects Stock Price Direction Using Machine Learning Model 4.1 Background of the Study 4.2 Machine Learning Models Used in This Case Study 4.3 Experiment Design and Evaluation 4.4 Analysis and Discussion of Results 5 Conclusion References Recommendations of Stockbrokers Versus Fuzzy Portfolio Approach in Construction Sector 1 Introduction 2 Fuzzy Numbers—Overview 2.1 Oriented Fuzzy Numbers 2.2 Trapezoidal-Oriented Fuzzy Numbers 2.3 Energy and Entropy Measures 3 Methods 3.1 Oriented Present Value 3.2 Expected Discount Factor 3.3 Portfolio Approach 4 Findings 4.1 Case Study—Portfolio π1 4.2 Case Study—Portfolio π2 5 Conclusions References Time Series Analysis Applications Automatic Clustering for Seasonal Time Series Based on Entropy 1 Introduction 2 Literature Review 3 Methods 3.1 High-Dimensional Time Series (HDTS): Spatio-Temporal Data Pre-processing 4 Empirical Results 5 Discussions and Conclusions References Modelling and Predicting the Dynamics of Confirmed COVID-19 Cases Based on Climate Data 1 Introduction 2 The Data 3 The Model 4 Findings and Discussions 4.1 Estimated Models and Impact of Climate Variables 4.2 Forecasting Results 5 Conclusion References Least-Squares Wavelet Analysis of Rainfalls and Landslide Displacement Time Series Derived by PS-InSAR 1 Introduction 2 Materials and Methods 2.1 Study Region 2.2 Datasets 2.3 Pearson Correlation Method 2.4 Least-Squares Spectral Analysis (LSSA) 2.5 Least-Squares Cross-Spectral Analysis (LSCSA) 2.6 Least-Squares Wavelet Analysis (LSWA) 2.7 Least-Squares Cross-Wavelet Analysis (LSCWA) 3 Results 3.1 Results of Traditional Methods 3.2 The LSWAVE Results 4 Discussion 5 Conclusions References Time Series Forecasting Macroeconomic Forecasting Evaluation of MIDAS Models 1 Introduction 2 MIDAS Models Under Comparison 2.1 ADL-MIDAS Model 2.2 U-MIDAS Model 2.3 TF-MIDAS Model 3 Forecasting Performance Evaluation 3.1 Data Description 3.2 Evaluation Design 3.3 Unconditional Distribution of the Forecasting Errors 4 Forecasting Performance: A Meta-Regression Analysis 4.1 Description of the Meta-Regression Analysis 4.2 Meta-Regression Main Results 5 Conclusions References Relative Measures of Forecasting: Lambda-Family-Measures 1 Introduction 2 Forecast Errors: A Short Literature Review 2.1 Most Common Error Measures and Others in Forecasting 3 Theoretical Framework 3.1 Automatic Forecasting and Testing: Univariate Forecasting Techniques 4 An Empirical Application to Evaluate Automatic Forecasting Modelling of Tourism Data 5 Conclusions and Discussions Appendix References Recurrent Neural Networks for Forecasting Time Series with Multiple Seasonality: A Comparative Study 1 Introduction 2 Forecasting Problem and Data Representation 3 Recurrent Cells 3.1 LSTM 3.2 GRU 3.3 dLSTM 3.4 dRNNCell 3.5 adRNNCell 4 RNN Architecture 5 Experimental Study 6 Conclusion References Time Series Applications in Energy Markov Processes for the Management of a Microgrid 1 Introduction 2 System Modeling 2.1 Microgrid Structure 2.2 ARMA Model for Demand Prediction 2.3 Non-homogeneous Markov Model for Wind Generation 2.4 Homogeneous Markov Model for Photovoltaic Generation 3 Optimization Strategy 3.1 System Constraints 3.2 Optimization Process 4 Application 4.1 Scenarios Studied 4.2 Results 4.3 Conclusions References Deep Learning Applied to Wind Power Forecasting: A Spatio-Temporal Approach 1 Introduction 2 State of the Art 2.1 CNN 2.2 RNN/LSTM 2.3 Transformers 2.4 Metrics 2.5 Related Work 3 Methodology 3.1 Data Sources 3.2 Exploratory Analysis 3.3 Experimentation 4 Implemented Neural Architectures 4.1 LSTM 4.2 CNN 4.3 Combinations 4.4 ConvLSTM 4.5 ViT 5 Results 5.1 Benchmark Results 5.2 Degradation Study 6 Conclusion References Forecasting Natural Gas Prices with Spatio-Temporal Copula-Based Time Series Models 1 Introduction 2 Data Description 3 Statistical Methods 3.1 Copulas 3.2 Vine Copulas 3.3 Modeling Time Series with Spatio-Temporal Copulas 3.4 Emergence of Non-elliptical Probabilistic Forecasts 4 Forecasting Study 5 Conclusion References