ITISE 2021

Special Session Proposals

SS1. Continuous-Time Models: Theory and Applications

The use of dynamic stochastic models in different areas has been increasing in recent years. In particular, Continuous Time models arise naturally for describing physical/economics/social phenomena that are observed on irregularly spaced time grids. Theoretical advantages have not been yet fully exploited in practice. Indeed, in recent literature we still find applications mainly based on discrete-time models partially due to the well- established estimation procedures. This special session is devoted to the collection of the latest developments related to the class of Continuous-Time models. In particular, papers focused on but not limited to the following topics of interest are welcome:

  • Time-series forecasting model for Coronavirus outbreak and other epidemics
  • Estimation and Simulation Algorithms for Continuous Time Models.
  • New Continuous Time Models and related statistical properties of estimators.


Organizers:
Lorenzo Mercuri is an Associate Professor at the University of Milano. He holds a PhD in Mathematical Finance from the University of Milano-Bicocca. He is currently involved in an international research project financed by CREST Japan. His research interests include Stochastic processes, Option Pricing and Numerical Methods applied to Finance. Department of Economics, Management and Quantitative Methods, University of Milan, Via del Conservatorio 7 Milan, Italy.
Prof. Edit Rroji is an Assistant Professor at the University of Milano-Bicocca. She holds a PhD in Mathematical Finance from the University of Milano-Bicocca. Previously, she has worked at the University of Trieste and at Politecnico di Milano. Her research interests include the modeling of financial time series and actuarial mathematics. Department of Statistics and Quantitative Methods, University of Milano-Bicocca, Piazza dell’Ateneo Nuovo 1 Milan, Italy.


Special Session Proposals. Previous Edition (2021)

SS1. Forecasting COVID-19 Coronavirus Outbreak

Epidemic is a rapid and wide spread of infectious disease threatening many lives and economy damages. It is important to fore-tell the epidemic lifetime so to decide on timely and remedic actions. These measures include closing borders, schools, suspending community services and commuters. Resuming such curfews depends on the momentum of the outbreak and its rate of decay. Being able to accurately forecast the fate of an epidemic is an extremely important but difficult task. Due to limited knowledge of the novel disease, the high uncertainty involved and the complex societal-political factors that influence the widespread of the new virus, any forecast is anything but reliable. Another factor is the insufficient amount of available data. Data samples are often scarce when an epidemic just started. With only few training samples on hand, finding a forecasting model which offers forecast at the best efforts is a big challenge in machine learning. This section invites works and works-in-progress from both academia and industrial partners to share and present ideas which could contribute to understanding and hopefully subsiding this outbreak which has evolved to global pandemic.

Papers related but not limited to the following topics of interest are solicited:

  • Time-series forecasting model for Coronavirus outbreak and other epidemics
  • Machine learning, AI and Big Data for modeling epidemic outbreaks
  • Decision support tools and optimization for modelling and controlling epidemics
  • Hybrid models of forecasting and other states-of-arts for predicting epidemics
  • Social media and text mining for predicting lifetimes and trends of epidemics
  • Data analytics and correlations of past epidemics and Coronavirus outbreak
  • Forecasting of post-epidemic stock markets and worldwide economy impact
  • Other issues related to the epidemics: Government policy, medical resources, etc.


Organizers:
Prof. Simon James Fong, University of Macau, Macau SAR.
Prof. Nilanjan Dey, Techno India College of Technology, India
Prof. Rubén González Crespo, Universidad Internacional de La Rioja, Logroño, Spain
Prof. Enrique Herrera-Viedma, University of Granada, Spain
Prof. Antonio J. Tallón-Ballesteros, University of Huelva, Huelva, Spain


SS2. Computational Intelligence for Applied Time Series Forecasting in Complex Systems (CIATSFCS)

MOTIVATION

Over the past few decades, application of simple statistical procedures with considerable heuristic or judgmental input was the beginning of forecasting, then in the 80’s, sophisticated time series models started to be used by some of the dynamic system operators, and these approaches, were to become pioneering works in this field. Soft computing methods including support vectors regression (SVR), fuzzy inference system (FIS) and artificial neural networks (ANN) to time-series forecasting (TSF) has been growing rapidly in order to unify the field of forecasting and to bridge the gap between theory and practice, making forecasting useful and relevant for decision-making in many fields of the sciences. The purpose of this session is to hold smaller, informal meetings where experts in a particular field of forecasting can discuss forecasting problems, research, and solutions in the field of automatic control. There is generally a nominal registration fee associated with attendance. This session aims to debate in finding solutions for problems facing the field of forecasting. We wish to hear from people working in different research areas, practitioners, professionals and academicians involved in this problematic.

SCOPE

The session seeks to foster the presentation and discussion of innovative techniques, implementations and applications of different problems that are Forecasting involved, specially in real-world problems applied to control and automation.
• Time Series Analysis
• Time Series Forecasting
• Evaluation of Forecasting Methods and Approaches
• Forecasting Applications in Business, Energy and Price Demand, Hydrology, etc.
• Impact of Uncertainty on Decision Making
• Seasonal Adjustment
• Multivariate Time Series Modelling and Forecasting
• Marketing Forecasting
• Economic and Econometric Forecasting

  • Journal of Forecasting .

    • Dr.Cristian Rodriguez, the Guest Editor of the Special Issue on "Bayesian Time Series Forecasting" at the Journal of Forecasting and the organizer of the Special Session in ITISE-2022: "SS2. Computational Intelligence for Applied Time Series Forecasting in Complex Systems (CIATSFCS)".


    Organizers:
    Prof. Cristian Rodriguez Rivero,University of Amsterdam, c.m.rodriguezrivero@uva.nl, IEEE CIS, Co-Founder LA-CIS.
    Prof. Alvaro Orjuela Cañón,Universidad del Rods, dorjuelaco@ieee.org, IEEE CIS, Co-Founder Board of Directors of LA-CIS.
    Prof. Héctor Daniel Patiño,Universidad Nacional de San Juan, Argentine,dpatino@inaut.unsj.edu.ar. IEEE CIS.
    Prof. Julián Antonio Pucheta,Universidad Nacional de Córdoba, Argentine,jpucheta@efn.uncor.edu.
    Prof. Gustavo Juarez,Universidad Nacional de Tucumán, Argentine, juarez.gustavo@ieee.org, IEEE CIS.
    Prof. Leonardo Franco,School of Engineering in Informatics, University of Malaga, Spain. lfranco@lcc.uma.es. IEEE CIS.


    SS3. Power Systems, Renewable Energies and Smart Grids


    Organizer:
    Prof. Pitshou Bokoro,Head of Department: Electrical and Electronic Engineering Technology, University of Johannesburg


    SS4. Time series analysis in the statistical process control


    Control charts have gradually been approved in pioneer industries as effective tools used in statistical process control (SPC) to ensure quality and save manufacturing costs. It is mainly used to identify the change in the process before manufacturing nonconforming items in massive amounts. After introducing the basic theory of process monitoring by Shewhart, numerous control charts have been developed to achieve special objectives under various assumptions [1]. One of the main assumptions of statistical process monitoring (SPM) is that the sampled observations at different time points must be independent. Nevertheless, the independence assumption is not realistic from two types of practical experiences: (1) sampling in high frequency induce autocorrelation in some processes, and (2) sampling from processes, such as chemical and environmental that introduce inherent autocorrelation [2]. In fact, in some industrial/non-industrial processes (e.g., continuous manufacturing processes, financial processes, health care systems, environmental phenomena, network monitoring), a correlation exists among adjacent observations [3]. The autocorrelation, if ignored, can significantly influence the statistical features of traditional control charts. This has led to the extension of various charts for autocorrelated observations. Two model-based approaches can be used to treat the process when the serial correlation exists among observations. These approaches include the residual control charts and the modified control charts [4]. In the first approach, the control charts are applied to the residuals obtained after fitting a time series model to eliminate the correlated structure. The special cause chart (SCC) was developed by Alwan and Roberts [5] as an initial study in this approach. In the second approach, the correlated observations are directly used on the control charts in which the control limits are adjusted according to the autocorrelation structure. Vasilopoulos and Stamboulis [6] performed the initial study on proposing the modified control chart. In addition to the introduced approaches, neural network-based control charts can also be categorized as the third approach, in which the data are processed without the requirements of identifying models or making adjustments [7]. In the fourth approach, some sampling strategies are taken to reduce the effect of autocorrelation [8]. Among the recent studies using different approaches, the reader can refer to [9-12]. All studies on developing control charts for autocorrelated data seem to have their advantages and disadvantages. Therefore, it is necessary to present a more simple and effective SPC methodology for monitoring autocorrelated processes. Moreover, the existing approaches can be applied in practice. Environmental & social sciences, finance, renewable sources, manufacturing, medicine, agriculture, etc are among the attractive fields that experience violation of independence assumption in some cases. Such applications can be treated with the introduced approaches.

    Organizer:
    Prof. Alireza Faraz and Prof. Samrad Jafarian-Namin,Industrial Engineering Department Faculty of Engineering Yazd University, Yazd, Iran


    SS5. Forecasting in High Dimension and Complex/Big Data

    Forecasting high-dimensional data and analysis of time series with hundreds/thousand of attributesinations) is challenging problem, with application in multiple problems of real life (problems in economy, energy, climate, etc). Real problems in which you have to predict / analyze or treat large volumes of data are welcome to this session.


    Organizer:
    Prof. Luis Javier Herrera ,Dep. Computer Architecture and Computer Technology, University of Granada, Spain
    Prof. Fernando Rojas ,Dep. Computer Architecture and Computer Technology, University of Granada, Spain


    SS6. Computational Intelligence methods for Time Series

    Within the field of science and engineering, it is very common to have data arranged in the form of time series data which must be subsequently analyzed, modeled and classified with the eventual goal of predicting future values. The literature shows that all these tasks related to time series can be undertaken using computational intelligence methods. In fact, new and further computational intelligence approaches, their efficiency and their comparison to statistical methods and other fact-checked computational intelligence methods, is a significant topic in academic and professional projects and works. Therefore, this special session aims at showing to our research community high quality and state of the art computational intelligence (and statistical) related works, applied to time series data and their tasks: analysis, forecasting, classification, and clustering. Furthermore, the experts can, from the starting point that the works shown provide, discuss different solutions and research issues for these topics.


    Organizer:
    Prof. Héctor Pomares ,Dep. Computer Architecture and Computer Technology, University of Granada, Spain
    Prof. German Gutierrez,Dep. Computer Science, E.P.S. University Carlos III of Madrid, Spain


    SS7. COVID-19: Modelling Miss-reported data and Economic Impact.

    The present outbreak of COVID-19 disease, caused by the SARS-CoV-2 virus, has put the planet in quarantine. On January 30, 2020, the World Health Organization (WHO) declared the COVID-19 outbreak a "public health emergency of international concern”, and a pandemic on March 11. Due to the rapid spread of the disease and the changing nature of diagnosis protocols, the official counts worldwide severely miss-report the true number of infected individuals. This is a common feature of epidemiological data. On the other hand the extreme measures taken to curb the rate of infection have negatively impacted the economy. This session invites works reporting solutions to these two important problems related to COVID-19: what was not seen and how it has hurt our finances.


    Organizer:
    Prof. Argimiro Arratia ,Universitat Politécnica de Catalunya Dept. of Computer Science, Barcelona
    Alejandra Cabaña , Universitat Autònoma de Barcelona, Dept. de Matemàtiques.


    SS8. Forecasting electricity load and prices.


    Organizer:
    Prof. P. Smith ,The University of Sydney


    SS9. Health forecasting.

    Providing your health care was a new area of ​​supply and a valuable tool for preventing future health events or situations, such as: B. the leadership of the health service and the need for health care. It facilitates preventive medicine and intervention strategy in the health sector and informs in advance of the health operation so that it is possible to adopt adequate mitigation measures to the minimum risk and command management.


    Organizer:
    Prof. J. Wang ,The Monash University, Australia


    SS10. New Developments on Time-Series Modelling for Financial Data.

    This session aims at presenting the recent developments of Time Series Modelling applied to financial and energy futures data. In particular, a focus is on studies that develop and apply recent nonlinear econometric models to reproduce financial market dynamics, capture financial data properties (asymmetry, volatility clustering, kurtosis excess, nonmorality, etc.). Papers on high frequency data and nonparametric econometric models are particularly welcomed.


    Organizer:
    Prof. H. Zhag ,University of York, Uk


    SS11. Advances in hydro-meteorological time series analysis and forecasts.

    It is well-known that time series analysis is time-consuming when large data sets are used, soft computing methods being recommended for obtaining a balance between the models' accuracy and speed of solving the problem at hand. Therefore, this special session aims to present the advances in the fields of modeling the hydro-meteorological time series and water quality assessment. Submissions reflecting theoretical methods and experimental works in the field of statistical analysis and applications to modeling hydro-meteorological time series, water quality assessment are expected.

    Suggested topics of this special session include but are not limited to: o Parametrical versus non-parametric approaches in hydro-meteorological data modeling o Critical evaluation and comparisons of alternative approaches for hydro-meteorological series modeling o New techniques for spatial data analysis applied to hydro-meteorology o New software for data analysis – development and applications to hydro-meteorological time series o Soft computing and fuzzy techniques in water hydro-meteorological time series modeling o Forecasting hydro-meteorological series o New techniques for water quality evaluation, monitoring, and forecast


    Organizer:
    Prof. dr. habil. Alina Bărbulescu , Technical University of Civil Engineering, Bucharest

  • MDPI Journal: WATER .

    • Prof. dr. habil. Alina Barbulescu, the Guest Editor of the Special Issue on "Assessing Hydrological Drought in a Climate Change: Methods and Measures" at the Journal WATER and the organizer of the Special Session in ITISE-2022: "SS11. Advances in hydro-meteorological time series analysis and forecasts.".


    SS12. Time Series for Natural Hazard Monitoring.


    Organizer:
    Prof. Dr. Dwivedi, R, , M N Natl Inst Technol Allahabad, Geog Informat Syst Cell, Allahabad 211004, Uttar Pradesh, India.


    SS13. Causal Reasoning Modelling for hydrological and multipurpose forecasting.

    Causal Reasoning (CR) should be seen as a reasoning pattern whose main goal is to predict the consequences or effects of some previous factors (Pearl, J., 2009). For instance, a joint distribution PB specifies the probability PB (A = a |E = e) of any event a given any observations e. The probability of the event a is computed by summing the probabilities of all of the entries in the resulting posterior distribution that are consistent with a. Queries such as these, where it is about the prediction of “downstream” effects of various factors, are instances of causal reasoning or prediction.

    Causality in hydrological records has not been deeply studied and it could be done by means of the joint use of different forms of reasoning patterns. These forms are Causal Reasoning (CR), Evidential Reasoning (ER) and Intercausal Reasoning (IR) (Koller and Friedman, 2009; Pearl, J., 2009). CR is used when the approach is done from top to bottom. In this sense, the analysis is focused on the cause and the objective comprises the prediction of the effect or consequence. Consequently, the queries in form of conditional probability, where the “downstream” effects of various factors are predicted, are instances of causal reasoning or prediction. ER comprises bottom-up reasoning, so the analysis is focused on the consequence (effect) and the cause is inferred (Bayesian Inference). IR is probably the hardest concept to understand. It comprises the interaction of different causes for the same effect. This type of reasoning is very useful in Hydrology, where a consequence can be generated or explained from several causes. Furthermore, one of the most exciting prospects in recent years has been the possibility of using the theory of Bayesian Networks to discover causal structures in raw data (Historical runoff record) (Pearl, J., 2014). This is performed through the usage of historical runoff data to train and populate the BN implementation. Consequently, AI techniques such as CR and ER and/or IR provide new horizons for this type of studies

    Furthermore, temporal dependence of hydrological time series has been deeply studied through classic and new approaches (Hao and Singh, 2016; Mishra and Singh, 2010; Mishra and Singh, 2011; Molina et al., 2016; Molina and Zazo, 2017). Conversely, spatial and spatio-temporal dependence for hydrologic science and engineering is much poorer studied (Holmström et al., 2015; Macián-Sorribes et al., 2020) and even more through Bayesian approaches (Lasinio et al., 2007; Wikle et al., 1998). This is because of some reasons explained as follows: complexity of characterizing and differentiating water sub-systems, scarcity of spatial data availability, difficulties in the application of spatial statistical methods, among others. Consequently, there is a general clear necessity of strengthening the spatio-temporal dependence studies on water systems (Holmström et al., 2015) and multipurposes through Causal Reasoning Modelling.


    Organizers:
    Dr. José-Luis Molina , IGA Research Group (University of Salamanca).
    Dr. Santiago Zazo
    Dr. Ana María Martín Casado
    Dra. Carmen Patino
    Dr. Fernando Espejo
    D. Abedin Hosseinpour



    SS14. Theoretical and Methodological Advances in High-Dimensional Time Series Analysis.

    Recent years have seen a rapid increase in the availability of high-dimensional time series in diverse contexts such as web traffic, sensor networks, finance, econometrics, neuroimaging, functional genomics, and more. To facilitate theory and computations, statistical methods for high-dimensional time series crucially rely on the concepts of sparsity, parsimony, and dimension reduction. Regularization methods based on LASSO, SCAD, and MCP penalties, for example, are widely used to induce sparsity in high-dimensional regression and covariance estimation. Dynamic factor models constitute a popular approach to reducing the dimension of time series. Nonstationarity, a problem that often compounds the high dimensionality of time series, can be tackled with locally stationary models, regime-switching models, latent process models, and segmentation methods.

    This session aims to attract novel theoretical and methodological contributions to the analysis of high-dimensional time series. In addition to the aforementioned topics, themes of interest include but are not restricted to: prediction, forecasting, variable selection, classification, change point detection, low rank + sparse methods, and spectral domain analysis.


    Organizer:
    Dr. David Degras , Assistant Professor, Department of Mathematics, University of Massachusetts Boston, web: um-boston.academia.edu/DavidDegras
    David Degras received his PhD in Statistics from the Université Paris 6, France, in 2008. He was a Postdoctoral Researcher at the Statistical and Applied Mathematical Sciences Institute (SAMSI) in 2010-11 and served as an Assistant Professor in the Department of Mathematical Sciences at DePaul University from 2011 to 2016. He is currently an Assistant Professor in the Department of Mathematics at the University of Massachusetts Boston. His research interests include computational statistics, convex and combinatorial optimization, neuroimaging, statistical learning, and functional data analysis.