Engineering Proceedings

Editorial

Jump to: Other

1 pages, 172 KiB

Open AccessEditorial

Statement of Peer Review

by Olga Valenzuela, Fernando Rojas, Luis Javier Herrera, Hector Pomares and Ignacio Rojas

Eng. Proc. 2022, 18(1), 43; https://doi.org/10.3390/engproc2022018043 - 6 Sep 2022

Viewed by 964

Abstract

In submitting conference proceedings to Engineering Proceedings, the volume editors of the proceedings certify to the publisher that all papers published in this volume have been subjected to peer review performed by the volume editors [...] Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

Other

Jump to: Editorial

10 pages, 436 KiB

Open AccessProceeding Paper

Evaluating a Recurrent Neural Network Model for Predicting Readmission to Cardiovascular ICUs Based on Clinical Time Series Data

by Sobhan Moazemi, Sebastian Kalkhoff, Steven Kessler, Zeynep Boztoprak, Vincent Hettlich, Artur Liebrecht, Roman Bibo, Bastian Dewitz, Artur Lichtenberg, Hug Aubin and Falko Schmid

Eng. Proc. 2022, 18(1), 1; https://doi.org/10.3390/engproc2022018001 - 16 Jun 2022

Cited by 2 | Viewed by 1469

Abstract

Unexpected readmission to intensive care units (ICUs) endangers patients’ lives due to premature patient transfers or prolonged stays at the care units. This can be mitigated by stratification of the readmission risk at discharge times using state-of-the-art machine learning (ML) methods. We fitted [...] Read more.

Unexpected readmission to intensive care units (ICUs) endangers patients’ lives due to premature patient transfers or prolonged stays at the care units. This can be mitigated by stratification of the readmission risk at discharge times using state-of-the-art machine learning (ML) methods. We fitted two alternative recurrent neural network (RNN) models based on long short-term memory (LSTM) on the Medical Information Mart for Intensive Care (MIMIC-III) dataset and evaluated them with an independent cohort from our hospital’s ICU (UKD). The first model processed all the available time series data from each patient’s ICU stay, whereas the second model focused on the data from the last 48 hours of the ICU stay prior to transfer. Our readmission prediction on MIMIC data reached an area under the curve of receiver operating characteristic (AUC-ROC) of 0.82. Furthermore, the model with the 48 h time frame outperformed the other model, as both models were applied to the independent test cohort. The results suggest that the RNN model for time series forecasting holds promise for future use as a clinical decision support tool, although follow-up studies with larger cohorts as well as user studies should be conducted to assess the generalizability and usability of the methods, respectively. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 550 KiB

Open AccessProceeding Paper

K-Means Clustering Assisted Spectrum Utilization Prediction with Deep Learning Models

by Bethelhem S. Shawel, Frehiwot Bantigegn, Tsegamlak T. Debella, Sofie Pollin and Dereje H. Woldegebreal

Eng. Proc. 2022, 18(1), 2; https://doi.org/10.3390/engproc2022018002 - 17 Jun 2022

Viewed by 1680

Abstract

The radio spectrum is a finite and scarce resource needed to transport data generated by existing and emerging wireless mobile networks and services. As the demand for wireless services is increasing, operators look for ways to efficiently utilize their assigned spectrum. While operators [...] Read more.

The radio spectrum is a finite and scarce resource needed to transport data generated by existing and emerging wireless mobile networks and services. As the demand for wireless services is increasing, operators look for ways to efficiently utilize their assigned spectrum. While operators do regularly perform spectrum occupancy measurement using an external spectrum analyzer or installing a dedicated sensing network to understand and plan the spectrum utilization level, both in time and spatial dimensions, such a measurement-based approach is expensive, given the dynamic and wide area covered by spectrum utilization. This paper proposes an indirect approach to assess and predict the average spectrum utilization level using data traffic measured from base stations of an operator network. K-Means clustering and deep learning algorithms, namely Convolution Neural Network (CNN) and Long Short Term Memory (LSTM), are used to model and analyze the current and future spectrum utilization in the 900 MHz frequency range. Data collected from 639 base stations of a mobile operator are used to build the spectrum utilization model. The results show that the CNN model trained on clustered data outperforms the model developed on non-clustered data (with a Root Mean Square Error (RMSE) of 0.58), mainly for base station level prediction. In terms of utilization level, the results also show that the operator does not optimally utilize the 900 MHz range. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

12 pages, 880 KiB

Open AccessProceeding Paper

Alone We Can Do So Little; Together We Cannot Be Detected

by Sergej Korlakov, Gerhard Klassen, Marcus Bravidor and Stefan Conrad

Eng. Proc. 2022, 18(1), 3; https://doi.org/10.3390/engproc2022018003 - 17 Jun 2022

Viewed by 1292

Abstract

It is no longer possible to imagine our everyday life without time series data. This includes, for example, market developments, COVID-19 cases, electricity prices, and other data from a wide variety of domains. An important task in the analysis of these data is [...] Read more.

It is no longer possible to imagine our everyday life without time series data. This includes, for example, market developments, COVID-19 cases, electricity prices, and other data from a wide variety of domains. An important task in the analysis of these data is the detection of anomalies. In most cases, this is accomplished by examining individual time series. In our work, we use the techniques of cluster analysis to establish a relationship between time series and groups of time series. This relationship allows us to observe the development of time series in their entirety, thereby gaining additional insights. Our approach identifies outliers with a real-world reference and enables the user to locate outliers without prior knowledge. To underline the strengths of our approach, we compare our method with another known method on two real-world datasets. We found that our solution needs significantly fewer calculations, produces more reasonable results, and can be applied to real-time data. Moreover, our method detected additional outliers, whose occurrence could be explained by real events. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 477 KiB

Open AccessProceeding Paper

ODIN TS: A Tool for the Black-Box Evaluation of Time Series Analytics

by Niccolò Zangrando, Rocio Nahime Torres, Federico Milani and Piero Fraternali

Eng. Proc. 2022, 18(1), 4; https://doi.org/10.3390/engproc2022018004 - 20 Jun 2022

Cited by 1 | Viewed by 990

Abstract

The increasing availability of time series datasets enabled by the diffusion of IoT architectures and the progress in the analysis of temporal data fostered by Deep Learning methods are boosting the interest in anomaly detection and predictive maintenance applications. The analysis of performance [...] Read more.

The increasing availability of time series datasets enabled by the diffusion of IoT architectures and the progress in the analysis of temporal data fostered by Deep Learning methods are boosting the interest in anomaly detection and predictive maintenance applications. The analysis of performance for these tasks relies on standard metrics applied to the entire dataset. Such indicators provide a global performance assessment but might not provide a deep understanding of the model weaknesses. A complementary diagnostic approach exploits error categorization and ad hoc visualizations. In this paper, we present ODIN TS, an open source diagnosis framework for time series analysis that lets developers compute performance metrics, disaggregated by different criteria, and visualize diagnosis reports. ODIN TS is agnostic to the training platform and can be extended with application- and domain-specific meta-annotations and metrics with almost no coding. We show ODIN TS at work through two time series analytics examples. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 365 KiB

Open AccessProceeding Paper

Cloud-Base Height Estimation Based on CNN and All Sky Images

by Emanuele Ogliari, Alfredo Nespoli, Elena Collino and Dario Ronzio

Eng. Proc. 2022, 18(1), 5; https://doi.org/10.3390/engproc2022018005 - 20 Jun 2022

Cited by 2 | Viewed by 1173

Abstract

Among several meteorological parameters, Cloud-Base Height is employed in many applications to provide operational and real-time cloud-base information to the aviation industry, to initialize Numeric Weather Prediction models and to validate climate models. Moreover, Cloud-Base Height is also useful in the nowcasting (very [...] Read more.

Among several meteorological parameters, Cloud-Base Height is employed in many applications to provide operational and real-time cloud-base information to the aviation industry, to initialize Numeric Weather Prediction models and to validate climate models. Moreover, Cloud-Base Height is also useful in the nowcasting (very short-term forecasting) of solar radiation. As cloud movements mainly affect the solar irradiance availability, their characterization is extremely important for solar power applications; an accurate estimation of the ground shadowing requires the knowledge of cloud height and extent. In the present work, the Cloud-Base Height value is estimated starting from sky images acquired from a single All Sky Imager. In order to fulfill this task, a Convolutional Neural Network model is chosen and developed. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 2070 KiB

Open AccessProceeding Paper

A Hybrid Model of VAR-DCC-GARCH and Wavelet Analysis for Forecasting Volatility

by Maryam Nafisi-Moghadam and Shahram Fattahi

Eng. Proc. 2022, 18(1), 6; https://doi.org/10.3390/engproc2022018006 - 20 Jun 2022

Cited by 2 | Viewed by 2115

Abstract

The purpose of this study is to investigate the time-varying co-movement between the volatility of gold, exchange rate, and stock market returns in Iran, using weekly data from 27 September 2013 to 3 December 2021. The results of the wavelet-based random forest show [...] Read more.

The purpose of this study is to investigate the time-varying co-movement between the volatility of gold, exchange rate, and stock market returns in Iran, using weekly data from 27 September 2013 to 3 December 2021. The results of the wavelet-based random forest show that the performance of VAR-DCC-GARCH model is better than that of DCC-GARCH model in predicting financial market volatilities. Furthermore, the results of the VAR-DCC-GARCH model indicate that a positive and relatively high conditional correlation exists in the daily exchange rate and gold-return volatility. The conditional correlation is lower between the exchange rate–stock market returns and the daily gold–stock market returns. In short and long terms, there is no correlation between the exchange rate and the stock market volatilities as well as gold and stock market volatilities, while the correlation between the paired markets exists in the medium term. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 1700 KiB

Open AccessProceeding Paper

Synthetic Subject Generation with Coupled Coherent Time Series Data

by Xabat Larrea, Mikel Hernandez, Gorka Epelde, Andoni Beristain, Cristina Molina, Ane Alberdi, Debbie Rankin, Panagiotis Bamidis and Evdokimos Konstantinidis

Eng. Proc. 2022, 18(1), 7; https://doi.org/10.3390/engproc2022018007 - 21 Jun 2022

Cited by 5 | Viewed by 1692

Abstract

A large amount of health and well-being data is collected daily, but little of it reaches its research potential because personal data privacy needs to be protected as an individual’s right, as reflected in the data protection regulations. Moreover, the data that do [...] Read more.

A large amount of health and well-being data is collected daily, but little of it reaches its research potential because personal data privacy needs to be protected as an individual’s right, as reflected in the data protection regulations. Moreover, the data that do reach the public domain will typically have under-gone anonymization, a process that can result in a loss of information and, consequently, research potential. Lately, synthetic data generation, which mimics the statistics and patterns of the original, real data on which it is based, has been presented as an alternative to data anonymization. As the data collected from health and well-being activities often have a temporal nature, these data tend to be time series data. The synthetic generation of this type of data has already been analyzed in different studies. However, in the healthcare context, time series data have reduced research potential without the subjects’ metadata, which are essential to explain the temporal data. Therefore, in this work, the option to generate synthetic subjects using both time series data and subject metadata has been analyzed. Two approaches for generating synthetic subjects are proposed. Real time series data are used in the first approach, while in the second approach, time series data are synthetically generated. Furthermore, the first proposed approach is implemented and evaluated. The generation of synthetic subjects with real time series data has been demonstrated to be functional, whilst the generation of synthetic subjects with synthetic time series data requires further improvements to demonstrate its viability. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 544 KiB

Open AccessProceeding Paper

Price Dynamics and Measuring the Contagion between Brent Crude and Heating Oil (US-Diesel) Pre and Post COVID-19 Outbreak

by Claudio Marcio Cassela Inacio, Jr. and Sergio Adriani David

Eng. Proc. 2022, 18(1), 8; https://doi.org/10.3390/engproc2022018008 - 21 Jun 2022

Cited by 4 | Viewed by 844

Abstract

The objective of this work is to analyze the price dynamics and the level of association between the Brent crude oil prices and heating oil (HO), i.e., US diesel. The data series are obtained from daily future contract prices of Chicago Mercantile Exchange [...] Read more.

The objective of this work is to analyze the price dynamics and the level of association between the Brent crude oil prices and heating oil (HO), i.e., US diesel. The data series are obtained from daily future contract prices of Chicago Mercantile Exchange (CME) group exchanges and the Intercontinental Exchange (ICE). A continuous evaluation of the Detrended Cross-Correlation Analysis (DCCA) between Brent crude oil prices vis-a-vis HO is proposed by means of the rolling window approach, allowing a dynamic analysis of their cross-correlations covering two periods, namely from January 2018 to December 2019 (before the COVID-19 pandemic) and from January 2020 to December 2021 (during the COVID-19 pandemic). The results indicate that there is a strong evidence of contagion in cross-correlation due to the initial impact of the pandemic, but the HO–Brent correlation fully recovered after approximately 200 days. However, lower time scales (n) are also sensitive to supply shortages in the short term and can be most reliable for agents that might not take long positions. Measuring this dynamic cross-correlation can provide useful information for investors and agents in the oil and energy markets. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 845 KiB

Open AccessProceeding Paper

Hybrid K-Mean Clustering and Markov Chain for Mobile Network Accessibility and Retainability Prediction

by Amel Salem Omer, Tesfaye Addisie Yemer and Dereje Hailemariam Woldegebreal

Eng. Proc. 2022, 18(1), 9; https://doi.org/10.3390/engproc2022018009 - 21 Jun 2022

Viewed by 1024

Abstract

To provide reliable services, mobile network operators (MNOs) continuously collect vital mobile network performance data to monitor and analyze the functioning of their radio access networks (RANs). RAN is a critical infrastructure for mobile networks and its performance is measured by key performance [...] Read more.

To provide reliable services, mobile network operators (MNOs) continuously collect vital mobile network performance data to monitor and analyze the functioning of their radio access networks (RANs). RAN is a critical infrastructure for mobile networks and its performance is measured by key performance indicators (KPIs) such as accessibility, retainability, availability, integrity, and mobility. The standard practice is that network managers utilize KPIs to identify failures or unusual events that can significantly degrade the quality of service delivery and the end-users’ experiences. However, taking corrective steps based on monitored performance parameters is a reactive approach that contributes to network and service degradations until corrective actions are taken. With the monitoring and automation of RAN infrastructure performance in mind, this paper presents the Markov chain, a widely used probabilistic modeling approach, as a systematic method for jointly predicting network accessibility and retainability status, two of the crucial RAN performance measures. The novel joint prediction is proposed to have a single operation for both accessibility and retainability. Real-time hourly KPIs data was collected from 1530 cells (base stations) run by an operator’s network in Addis Ababa, Ethiopia, for 4 months, from 1 November 2020 to 28 February 2021. The cells are scattered across the capital city, where factors such as land use, settlement patterns, and customers behaviors differ. To capture the spatial variation of the KPIs without escalating the computational complexity much, the dataset is separated into six clusters using the K-mean clustering approach. The Markov chain KPIs status prediction models are formulated on a cluster level. The results reveal that the proposed models can predict the KPIs status with 94.61 percent accuracy. Because the data is already available and can be collected at any time using the operator’s network management system (NMS), this is a cost-effective technique to proactively improve mobile network performance. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 3529 KiB

Open AccessProceeding Paper

A Multivariate Approach for Spatiotemporal Mobile Data Traffic Prediction

by Bethelhem S. Shawel, Endale Mare, Tsegamlak T. Debella, Sofie Pollin and Dereje H. Woldegebreal

Eng. Proc. 2022, 18(1), 10; https://doi.org/10.3390/engproc2022018010 - 21 Jun 2022

Cited by 4 | Viewed by 1769

Abstract

Widespread deployment of spectrally efficient mobile networks, advancements in mobile devices, and proliferation of attractive applications has led to an exponential increase in mobile data traffic. Mobile Network Operators (MNOs) benefit from the associated revenue generation while putting efforts to meet customers’ expectations [...] Read more.

Widespread deployment of spectrally efficient mobile networks, advancements in mobile devices, and proliferation of attractive applications has led to an exponential increase in mobile data traffic. Mobile Network Operators (MNOs) benefit from the associated revenue generation while putting efforts to meet customers’ expectations of delivered services. Having a clear knowledge of the traffic demand is critical for network dimensioning, optimization, resource allocation, market planning, and the like. As the traffic demand, among others, is a function of customers’ behavior and settlement patterns, land use, and time of the day, capturing traffic characteristics in both temporal and spatial dimensions is needed. Moreover, other parameters, such as the number of users and data throughput, inherently contain traffic-related information, necessitating a multivariate approach for understanding the traffic demand. Realizing the multidimensional and multivariate nature of the mobile data traffic, in this paper, we propose a multivariate and hybrid Convolutional Neural Network and Long Short-Term Memory network (CNN-LSTM) data traffic prediction model. The model is built on mobile traffic data collected from a Network Operator for Long-Term Evolution (LTE) network. The results confirm that the proposed model outperforms its univariate counterparts in Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) by 58% and 50%, respectively. Moreover, the model is further compared with CNN-only univariate and multivariate models, which it also outperforms. The comparisons substantiate the achievable improvements because of the hybrid and multivariate nature of the prediction algorithm. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 951 KiB

Open AccessProceeding Paper

An Application of Neural Networks to Predict COVID-19 Cases in Italy

by Lorena Saliaj and Eugenia Nissi

Eng. Proc. 2022, 18(1), 11; https://doi.org/10.3390/engproc2022018011 - 21 Jun 2022

Cited by 2 | Viewed by 922

Abstract

COVID-19 pandemic has become the greatest worldwide threat, as it has spread rapidly among individuals in most countries around the world. This study concerns the problem of weekly prediction of new COVID-19 cases in Italy, aiming to find the best predictive model for [...] Read more.

COVID-19 pandemic has become the greatest worldwide threat, as it has spread rapidly among individuals in most countries around the world. This study concerns the problem of weekly prediction of new COVID-19 cases in Italy, aiming to find the best predictive model for daily infection number in countries with a large number of confirmed cases. We compare the forecasting performance of linear and nonlinear forecasting models using weekly COVID-19 data for the period between 24 February 2020 until 16 May 2022. We discuss various forecasting approaches, including a Nonlinear Autoregressive Neural Network (NARNN) model, an Autoregressive Integrated Moving Average (ARIMA) model, a TBATS model, and Exponential Smoothing on the collected data and compared their accuracy using the data collected from 23 March 2020 to 20 April 2020, choosing the model with the lowest Mean Absolute Percentage Error (MAPE) value. Since the linear models seem to not easily follow the nonlinear patterns of daily confirmed COVID-19 cases, Artificial Neural Network (ANN) have been successfully applied to solve problems of forecasting nonlinear models. The model has been used for weekly prediction of COVID-19 cases for the next 4 weeks without any additional intervention. The prediction model can be applied to other countries struggling with the COVID-19 pandemic, to any possible future pandemics, and also help make better decisions in future. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 852 KiB

Open AccessProceeding Paper

Relationship between Stationarity and Dynamic Convergence of Time Series

by Gerardo Covarrubias and Xuedong Liu

Eng. Proc. 2022, 18(1), 12; https://doi.org/10.3390/engproc2022018012 - 21 Jun 2022

Viewed by 1214

Abstract

In economic science, when a stochastic process is studied from an econometric approach, it focuses on stationarity and the reference approach to dynamic equilibrium is ignored. Although both approaches are theoretically closely linked through the common purpose of making forecasts, in existing methodologies [...] Read more.

In economic science, when a stochastic process is studied from an econometric approach, it focuses on stationarity and the reference approach to dynamic equilibrium is ignored. Although both approaches are theoretically closely linked through the common purpose of making forecasts, in existing methodologies they are studied in a mutually exclusive way. Therefore, in this paper the consistency between dynamic equilibrium and stationarity is analyzed. In practice, the theoretical correspondence between these two important properties could present some inconsistency due to misspecification in an autoregressive model or the presence of spuriousness, generated by the components of the series. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

3 pages, 202 KiB

Open AccessProceeding Paper

Partitioning of Net Ecosystem Exchange Using Dynamic Mode Decomposition and Time Delay Embedding

by Maha Shadaydeh, Joachim Denzler and Mirco Migliavacca

Eng. Proc. 2022, 18(1), 13; https://doi.org/10.3390/engproc2022018013 - 21 Jun 2022

Viewed by 1041

Abstract

Ecosystem respiration (Reco) represents a major component of the global carbon cycle. An accurate estimation of Reco dynamics is necessary for a better understanding of ecosystem–climate interactions and the impact of climate extremes on ecosystems. This paper proposes a new data-driven method for [...] Read more.

Ecosystem respiration (Reco) represents a major component of the global carbon cycle. An accurate estimation of Reco dynamics is necessary for a better understanding of ecosystem–climate interactions and the impact of climate extremes on ecosystems. This paper proposes a new data-driven method for the estimation of the nonlinear dynamics of Reco using the method of dynamic mode decomposition with control input (DMDc). The method is validated on the half-hourly Fluxnet 2015 data. The model is first trained on the night-time net ecosystem exchange data. The day-time Reco values are then predicted using the obtained model with future values of a control input such as air temperature and soil water content. To deal with unobserved drivers of Reco other than the user control input, the method uses time-delay embedding of the history of Reco and the control input. Results indicate that, on the one hand, the prediction accuracy of Reco dynamics using DMDc is comparable to state-of-the-art deep learning-based methods, yet it has the advantages of being a simple and almost hyper-parameter-free method with a low computational load. On the other hand, the study of the impact of different control inputs on Reco dynamics showed that for most of the studied Fluxnet sites, air temperature is a better long-term predictor of Reco, while using soil water content as control input produced better short-term prediction accuracy. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

9 pages, 822 KiB

Open AccessProceeding Paper

An Ordinal Procedure to Detect Change Points in the Dependence Structure between Non-Stationary Time Series

by Alexander Schnurr and Svenja Fischer

Eng. Proc. 2022, 18(1), 14; https://doi.org/10.3390/engproc2022018014 - 21 Jun 2022

Cited by 1 | Viewed by 795

Abstract

The presence of non-stationarity can have crucial effects on statistical tests on correlation between two or more data sets. We present a procedure to detect changes in the strength of dependence between data sets that is based solely on a comparison of the [...] Read more.

The presence of non-stationarity can have crucial effects on statistical tests on correlation between two or more data sets. We present a procedure to detect changes in the strength of dependence between data sets that is based solely on a comparison of the ordinal structures within a moving window and thus compares the up-and-down behavior only. Hence, it is not distracted by changes within a single data set, such as change-points and trends or even nonlinear transformations, leading to non-stationarity. The applicability of the method is demonstrated for a hydrological data set of runoff time series which are impacted by a reservoir. It is demonstrated that the method overcomes problems of classical methods when non-stationarity is present. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 445 KiB

Open AccessProceeding Paper

On the Prospective Use of Deep Learning Systems for Earthquake Forecasting over Schumann Resonances Signals

by Carlos Cano-Domingo, Ruxandra Stoean, Nuria Novas-Castellano, Manuel Fernandez-Ros, Gonzalo Joya and Jose A. Gázquez-Parra

Eng. Proc. 2022, 18(1), 15; https://doi.org/10.3390/engproc2022018015 - 21 Jun 2022

Cited by 2 | Viewed by 1119

Abstract

The relationship between Schumann resonances and earthquakes was proposed more than 50 years ago; however, the experimental support has not been fully established. A considerable amount of recent studies have focused on the relationship between a single earthquake and the Schumann resonance signal [...] Read more.

The relationship between Schumann resonances and earthquakes was proposed more than 50 years ago; however, the experimental support has not been fully established. A considerable amount of recent studies have focused on the relationship between a single earthquake and the Schumann resonance signal variation around this earthquake, obtaining preliminary support for the existence of the link. Nonetheless, they all lack a systematic and general approach. In this research, we propose a novel methodology to detect the presence of relevant earthquakes based on the Schumann resonance. The methodology is based on a deep learning framework composed of a pretrained variational auto-encoder followed by an LSTM network and a fully connected layer with a sigmoid output. The results reveal the uncovered relationship between earthquake activity and Schumann resonance signal using the novel methodology, being the first automatic earthquake detector based on Schumann resonance signal. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 314 KiB

Open AccessProceeding Paper

Probabilistic Forecasting for Oil Producing Wells Using Seq2seq Augmented Model

by Hadeel Afifi, Mohamed Elmahdy, Motaz El Saban and Mervat Abu-Elkheir

Eng. Proc. 2022, 18(1), 16; https://doi.org/10.3390/engproc2022018016 - 21 Jun 2022

Cited by 1 | Viewed by 1003

Abstract

Time series forecasting is a challenging problem in the field of data mining. Deterministic forecasting has shown limitations in the field. Therefore, researchers are now more inclined towards probabilistic forecasting, which has shown a clear advantage by providing more reliable models. In this [...] Read more.

Time series forecasting is a challenging problem in the field of data mining. Deterministic forecasting has shown limitations in the field. Therefore, researchers are now more inclined towards probabilistic forecasting, which has shown a clear advantage by providing more reliable models. In this paper, we utilize seq2seq machine learning models in order to estimate prediction intervals (PIs) for a large oil production dataset. To evaluate the proposed models, Prediction Interval Coverage Probability (PICP), Prediction Interval Normalized Average Width (PINAW), and Coverage Width-based Criterion (CWC) metrics are used. Our results show that the proposed model can reliably estimate PIs for production forecasting. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 312 KiB

Open AccessProceeding Paper

Towards Time-Series Feature Engineering in Automated Machine Learning for Multi-Step-Ahead Forecasting

by Can Wang, Mitra Baratchi, Thomas Bäck, Holger H. Hoos, Steffen Limmer and Markus Olhofer

Eng. Proc. 2022, 18(1), 17; https://doi.org/10.3390/engproc2022018017 - 21 Jun 2022

Cited by 10 | Viewed by 2868

Abstract

Feature engineering is an essential step in the pipelines used for many machine learning tasks, including time-series forecasting. Although existing AutoML approaches partly automate feature engineering, they do not support specialised approaches for applications on time-series data such as multi-step forecasting. Multi-step forecasting [...] Read more.

Feature engineering is an essential step in the pipelines used for many machine learning tasks, including time-series forecasting. Although existing AutoML approaches partly automate feature engineering, they do not support specialised approaches for applications on time-series data such as multi-step forecasting. Multi-step forecasting is the task of predicting a sequence of values in a time-series. Two kinds of approaches are commonly used for multi-step forecasting. A typical approach is to apply one model to predict the value for the next time step. Then the model uses this predicted value as an input to forecast the value for the next time step. Another approach is to use multi-output models to make the predictions for multiple time steps of each time-series directly. In this work, we demonstrate how automated machine learning can be enhanced with feature engineering techniques for multi-step time-series forecasting. Specifically, we combine a state-of-the-art automated machine learning system, auto-sklearn, with tsfresh, a library for feature extraction from time-series. In addition to optimising machine learning pipelines, we propose to optimise the size of the window over which time-series data are used for predicting future time-steps. This is an essential hyperparameter in time-series forecasting. We propose and compare (i) auto-sklearn with automated window size selection, (ii) auto-sklearn with tsfresh features, and (iii) auto-sklearn with automated window size selection and tsfresh features. We evaluate these approaches with statistical techniques, machine learning techniques and state-of-the-art automated machine learning techniques, on a diverse set of benchmarks for multi-step time-series forecasting, covering 20 synthetic and real-world problems. Our empirical results indicate a significant potential for improving the accuracy of multi-step time-series forecasting by using automated machine learning in combination with automatically optimised feature extraction techniques. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

5 pages, 609 KiB

Open AccessProceeding Paper

PV Fault Diagnosis Method Based on Time Series Electrical Signal Analysis

by Carole Lebreton, Fabrice Kbidi, Frédéric Alicalapa, Michel Benne and Cédric Damour

Eng. Proc. 2022, 18(1), 18; https://doi.org/10.3390/engproc2022018018 - 21 Jun 2022

Cited by 1 | Viewed by 1273

Abstract

With the objectives of energy self-sufficiency and zero emissions in La Reunion, photovoltaics are becoming an increasingly important part of the local energy mix. Installation reliability and safety are crucial to ensure network stability and security, therefore PV system fault diagnosis is an [...] Read more.

With the objectives of energy self-sufficiency and zero emissions in La Reunion, photovoltaics are becoming an increasingly important part of the local energy mix. Installation reliability and safety are crucial to ensure network stability and security, therefore PV system fault diagnosis is an essential tool in the expansion of this electricity production method. The DETECT Project (Diagnosis onlinE of sTate of health of EleCTric systems) is a research project aiming at diagnosis method development. In this way, signal processing provides us with promising tools in the form of decomposition algorithms. Thanks to their low computation cost, empirical mode decomposition (EMD) and variational mode decomposition (VMD) allow undertaking a real-time diagnosis, with on-line PV electrical signal time-series data analysis. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 1168 KiB

Open AccessProceeding Paper

Early Detection of Flash Floods Using Case-Based Reasoning

by Enrique Fernádez, José R. Villar, Alberto Navarro and Javier Sedano

Eng. Proc. 2022, 18(1), 19; https://doi.org/10.3390/engproc2022018019 - 21 Jun 2022

Viewed by 888

Abstract

A flash flood is the sudden increase in the water level of a basin due to an abrupt change in weather conditions. The importance of early flash flood detection is given by reducing its consequences, either in infrastructure damage or human losses. Interestingly, [...] Read more.

A flash flood is the sudden increase in the water level of a basin due to an abrupt change in weather conditions. The importance of early flash flood detection is given by reducing its consequences, either in infrastructure damage or human losses. Interestingly, the studies in the literature focus on the dynamics of the basins, determining how the water levels in a basin would be in a considered scenario, and leaving the early and online flash flood detection unaddressed. This research addresses this latter problem and proposes a case-based reasoning that estimates the flooding map for a given prediction horizon. Provided enough data are available, this CBR tool would perfectly deal with different basins and locations. This research is being designed and developed on two concrete basins, one from Spain and one from France. We expect that the performance of the CBR tool will satisfactorily assess the decision making of the public safety experts. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

6 pages, 222 KiB

Open AccessProceeding Paper

Inland Areas, Protected Natural Areas and Sustainable Development

by Antonio Bertini, Immacolata Caruso and Tiziana Vitolo

Eng. Proc. 2022, 18(1), 20; https://doi.org/10.3390/engproc2022018020 - 21 Jun 2022

Cited by 4 | Viewed by 1002

Abstract

In recent years, policies implemented by many countries have resulted in the deterioration of the Earth’s environment rather than the protection of environmental resources. The impact on the environment and territories, particularly those in the Mediterranean basin, is now evident in all its [...] Read more.

In recent years, policies implemented by many countries have resulted in the deterioration of the Earth’s environment rather than the protection of environmental resources. The impact on the environment and territories, particularly those in the Mediterranean basin, is now evident in all its gravity. Even in Italy, the development policies pursued until 2010 favored urban settlements, neglecting the rest of the territory and, in particular, areas far from the traffic flows of people and goods. In the last decade, Italy has also begun to invest in ‘inland areas’ and protected natural areas that, if well analyzed, organized and managed, can become the promoter of sustainable development for the entire country. Starting from the potential expressed by local communities, landscape resources, cultural, intangible and tangible heritage, the aim is to provide direction and potential scenarios for enhancing economic opportunities, unexpressed and possible drivers of sustainable development. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

8 pages, 342 KiB

Open AccessProceeding Paper

Expectation-Maximization Algorithm for Autoregressive Models with Cauchy Innovations

by Monika S. Dhull and Arun Kumar

Eng. Proc. 2022, 18(1), 21; https://doi.org/10.3390/engproc2022018021 - 21 Jun 2022

Cited by 1 | Viewed by 1068

Abstract

In this paper, we study the autoregressive (AR) models with Cauchy distributed innovations. In the AR models, the response variable

y_{t}

depends on previous terms and a stochastic term (the innovation). In the classical version, the AR models are based on normal [...] Read more.

In this paper, we study the autoregressive (AR) models with Cauchy distributed innovations. In the AR models, the response variable

y_{t}

depends on previous terms and a stochastic term (the innovation). In the classical version, the AR models are based on normal distribution which could not capture the extreme values or asymmetric behavior of data. In this work, we consider the AR model with Cauchy innovations, which is a heavy-tailed distribution. We derive closed forms for the estimates of parameters of the considered model using the expectation-maximization (EM) algorithm. The efficacy of the estimation procedure is shown on the simulated data. The comparison of the proposed EM algorithm is shown with the maximum likelihood (ML) estimation method. Moreover, we also discuss the joint characteristic function of the AR(1) model with Cauchy innovations, which can also be used to estimate the parameters of the model using empirical characteristic function. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 1142 KiB

Open AccessProceeding Paper

Deep Representation Learning for Cluster-Level Time Series Forecasting

by Tsegamlak T. Debella, Bethelhem S. Shawel, Maxime Devanne, Jonathan Weber, Dereje H. Woldegebreal, Sofie Pollin and Germain Forestier

Eng. Proc. 2022, 18(1), 22; https://doi.org/10.3390/engproc2022018022 - 21 Jun 2022

Cited by 2 | Viewed by 1797

Abstract

In today’s data-driven world, time series forecasting is an intensively investigated temporal data mining technique. In practice, there is a range of forecasting techniques that have been proven to be efficient at capturing different aspects of an input. For instance, classic linear forecasting [...] Read more.

In today’s data-driven world, time series forecasting is an intensively investigated temporal data mining technique. In practice, there is a range of forecasting techniques that have been proven to be efficient at capturing different aspects of an input. For instance, classic linear forecasting models such as seasonal autoregressive integrated moving average (S-ARIMA) models are known to capture the trends and seasonality evident in temporal datasets. In contrast, neural-network-based forecasting approaches are known to be best at capturing nonlinearity. Despite such differences, most forecasting techniques inherently assume that models are fitted using a single input. In practice, there are often cases where we cannot deploy forecasting models in this manner. For instance, in most wireless communication traffic forecasting problems, temporal datasets are defined by taking samples from hundreds of base stations. Moreover, the base stations are expected to have spatial correlation due to user mobility, land use, settlement patterns, etc. Thus, in such cases, it is often advised that forecasting should be approached using clusters that group the base stations based on their traffic patterns. However, when this approach is used, the quality of the cluster centroids and the overall cluster formation process is expected to have a significant impact on the performance of forecasting models. In this paper, we show the effectiveness of representation learning for cluster formation and cluster centroid definition, which in turn improves the quality of cluster-level forecasting. We demonstrate this concept using data traffics collected from 729 wireless base stations. In general, based on the experimental results, the representation learning approach outperforms cluster-level forecasting models based on classical clustering techniques such as K-means and dynamic time warping barycenter averaging K-means (DBA K-means). Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 4138 KiB

Open AccessProceeding Paper

Autoencoders for Anomaly Detection in an Industrial Multivariate Time Series Dataset

by Theodoros Tziolas, Konstantinos Papageorgiou, Theodosios Theodosiou, Elpiniki Papageorgiou, Theofilos Mastos and Angelos Papadopoulos

Eng. Proc. 2022, 18(1), 23; https://doi.org/10.3390/engproc2022018023 - 22 Jun 2022

Cited by 18 | Viewed by 4783

Abstract

In smart manufacturing, the automation of anomaly detection is essential for increasing productivity. Timeseries data from production processes are often complex sequences and their assessment involves many variables. Thus, anomaly detection with deep learning approaches is considered as an efficient and effective methodology. [...] Read more.

In smart manufacturing, the automation of anomaly detection is essential for increasing productivity. Timeseries data from production processes are often complex sequences and their assessment involves many variables. Thus, anomaly detection with deep learning approaches is considered as an efficient and effective methodology. In this work, anomaly detection with deep autoencoders is examined. Three autoencoders are employed to analyze an industrial dataset and their performance is assessed. Autoencoders based on long short-term memory and convolutional neural networks appear to be the most promising. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 2411 KiB

Open AccessProceeding Paper

Time Series Clustering of High Gamma Dose Rate Incidents

by Mohammed Al Saleh, Beatrice Finance, Yehia Taher, Ali Jaber and Roger Luff

Eng. Proc. 2022, 18(1), 24; https://doi.org/10.3390/engproc2022018024 - 22 Jun 2022

Viewed by 1083

Abstract

In this paper, we proposed an unsupervised machine-learning-based framework to automate the process of extracting suspicious gamma dose rate incidents from the real unlabeled raw historical data measured in the German Radiation Early Warning Network and identify the underlying events behind each. This [...] Read more.

In this paper, we proposed an unsupervised machine-learning-based framework to automate the process of extracting suspicious gamma dose rate incidents from the real unlabeled raw historical data measured in the German Radiation Early Warning Network and identify the underlying events behind each. This raised the research problem of clustering unlabeled time series data with varying lengths and scales. Based on the many evaluations, we demonstrated that the state-of-the-art’s most popular time series clustering models were not suitable to perform this task. This motivated us to introduce our own approach. Through this approach we were able to perform online classification for gamma dose rate incidents of varying lengths and scales. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 284 KiB

Open AccessProceeding Paper

A Dynamic Combination of Theta Method and ATA: Validating on a Real Business Case

by Yasin Tadayonrad and Alassane Ballé Ndiaye

Eng. Proc. 2022, 18(1), 25; https://doi.org/10.3390/engproc2022018025 - 22 Jun 2022

Viewed by 919

Abstract

In order to make better decisions and take efficient actions in any supply chain system, we need to have better estimation of uncertain parameters, especially the future demands of our customers. To do so we must use a forecasting model which gives the [...] Read more.

In order to make better decisions and take efficient actions in any supply chain system, we need to have better estimation of uncertain parameters, especially the future demands of our customers. To do so we must use a forecasting model which gives the most useful and accurate forecasts. Time series forecasting methods are still one of the most popular approaches used in the business because of their simplicity. One of the most recent methods that caught the attention of researchers and practitioners is Theta method, which was first ranked in M3 competition. This method works based on the decomposition of the deseasonalized original demand data into two components. The first component represents the long-term trend, and the second component indicates the short-term behavior of the data set. ATA method is another method which has been introduced recently. ATA method works like exponential smoothing methods, but in ATA method the smoothing parameter is a function of time point. In this paper we have proposed a new form of Theta method in which we have benefited from the features of ATA and presented a combination of ATA method and Theta method. We have introduced a dynamic model which uses Theta method as the main model and selected from among some alternative methods such as ATA method, simple exponential, and Double Exponential smoothing methods to be used as the theta lines. Also, we optimize the parameters of each method used in the model. Finally, we have tested the mentioned model on a real data set and concluded that the combination of Theta and ATA methods has a better performance compared to the other alternatives in terms of forecast accuracy. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

7 pages, 1068 KiB

Open AccessProceeding Paper

Limitation of Deep-Learning Algorithm for Prediction of Power Consumption

by Majdi Frikha, Khaled Taouil, Ahmed Fakhfakh and Faouzi Derbel

Eng. Proc. 2022, 18(1), 26; https://doi.org/10.3390/engproc2022018026 - 22 Jun 2022

Cited by 4 | Viewed by 1606

Abstract

In recent years, electricity consumption has become high due to the use of several domestic applications in the house. On the other hand, there is a trend of using renewable energy in many houses, such as solar energy, energy-storage systems and electric vehicles. [...] Read more.

In recent years, electricity consumption has become high due to the use of several domestic applications in the house. On the other hand, there is a trend of using renewable energy in many houses, such as solar energy, energy-storage systems and electric vehicles. For this reason, forecasting household electricity consumption is essential for managing and planning energy use. Forecasting power consumption is a difficult time-series-forecasting task. Additionally, the electrical load has irregular trend elements, which makes it very difficult to predict the demand for electrical energy using simple forecasting techniques. Therefore, several researchers have worked on intelligent algorithms such as machine-learning and deep-learning algorithms to find a solution for this problem. In this work, we demonstrate that deep-learning algorithms are not always reliable and accurate in predicting power consumption. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 567 KiB

Open AccessProceeding Paper

Combination of Post-Processing Methods to Improve High-Resolution NWP Solar Irradiance Forecasts in French Guiana

by Rafael Alvarenga, Hubert Herbaux and Laurent Linguet

Eng. Proc. 2022, 18(1), 27; https://doi.org/10.3390/engproc2022018027 - 22 Jun 2022

Viewed by 1139

Abstract

Efforts have been made to improve Numerical Weather Prediction (NWP) forecasts using post-processing techniques, relying on statistical models to refine the weather forecasts. Most approaches used in the literature suffer from two main deficiencies when applied to high-resolution data: (1) they high capacity [...] Read more.

Efforts have been made to improve Numerical Weather Prediction (NWP) forecasts using post-processing techniques, relying on statistical models to refine the weather forecasts. Most approaches used in the literature suffer from two main deficiencies when applied to high-resolution data: (1) they high capacity models to retain nonlinear data fluctuations; (2) some are known to reduce the mean random error; however, they may still generate subsequent biased forecasts. In this study, methods from three different approaches are compared to improve 10-min resolution NWP solar irradiance forecasts, namely a neural network and a linear statistical model as Model Output Statistics, Kalman Filter and Kernel Conditional Density Estimation. The results show that none of the methods, if used individually, improve the mean absolute error (MAE) and mean bias (MBE) jointly. However, a combination of a neural network followed by Kalman filter post-processing results in significant improvements both in the mean random error and the systematic mean bias of original forecasts, reducing the MAE by 45% and the MBE by 91%, respectively. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 1191 KiB

Open AccessProceeding Paper

Online Classification of High Gamma Dose Rate Incidents

by Mohammed Al Saleh, Beatrice Finance, Yehia Taher, Ali Jaber and Roger Luff

Eng. Proc. 2022, 18(1), 28; https://doi.org/10.3390/engproc2022018028 - 22 Jun 2022

Viewed by 891

Abstract

In this paper, we propose a new method for choosing the most suitable time-series classification method that can be applied to online gamma dose rate incidents. We referred to the historical incidents measured in the German Radiation Early Warning Network and clustered them [...] Read more.

In this paper, we propose a new method for choosing the most suitable time-series classification method that can be applied to online gamma dose rate incidents. We referred to the historical incidents measured in the German Radiation Early Warning Network and clustered them into several classes before testing existing classification methods. This raises the research problem of the online classification of time-series data with varying scales and lengths. Referring to the state-of-the-art methods, we found that no specific classification method can fit our data all the time. This motivated us to introduce our own approach. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 2810 KiB

Open AccessProceeding Paper

Comparative Analysis of Residential Load Forecasting with Different Levels of Aggregation

by Ana Apolo Peñaloza, Roberto Chouhy Leborgne and Alexandre Balbinot

Eng. Proc. 2022, 18(1), 29; https://doi.org/10.3390/engproc2022018029 - 21 Jun 2022

Cited by 2 | Viewed by 1074

Abstract

Microgrids need a robust residential load forecasting. As a consequence, this highlights the problem of predicting electricity consumption in small amounts of households. The individual demand curve is volatile, and more difficult to forecast than the aggregated demand curve. For this reason, Mean [...] Read more.

Microgrids need a robust residential load forecasting. As a consequence, this highlights the problem of predicting electricity consumption in small amounts of households. The individual demand curve is volatile, and more difficult to forecast than the aggregated demand curve. For this reason, Mean Absolute Percentage Error (MAPE) varies in a large range (of 1% to 45%), depending on the number of consumers analyzed. Different levels of aggregation of household consumers that can be used in microgrids are analyzed; the load forecasting of the single consumer and aggregated consumers are compared. The forecasting methodology used is the most consolidated of Recurrent Neural Networks, i.e., LSTM. The dataset used contains 920 residential consumers belonging to the Commission for Energy Regulation (CER), a control group that is in the Irish Social Science Data Archive (ISSDA) repository. The result shows that the forecasting of groups of more than 20 aggregated consumers has a lower MAPE that individual forecasting. On the other hand, individual forecasting is better for groups with fewer than 10 consumers. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 549 KiB

Open AccessProceeding Paper

An Open Source and Reproducible Implementation of LSTM and GRU Networks for Time Series Forecasting

by Gissel Velarde, Pedro Brañez, Alejandro Bueno, Rodrigo Heredia and Mateo Lopez-Ledezma

Eng. Proc. 2022, 18(1), 30; https://doi.org/10.3390/engproc2022018030 - 22 Jun 2022

Cited by 3 | Viewed by 2049

Abstract

This paper introduces an open source and reproducible implementation of Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks for time series forecasting. We evaluated LSTM and GRU networks because of their performance reported in related work. We describe our method and [...] Read more.

This paper introduces an open source and reproducible implementation of Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks for time series forecasting. We evaluated LSTM and GRU networks because of their performance reported in related work. We describe our method and its results on two datasets. The first dataset is the S&P BSE BANKEX, composed of stock time series (closing prices) of ten financial institutions. The second dataset, called Activities, comprises ten synthetic time series resembling weekly activities with five days of high activity and two days of low activity. We report Root Mean Squared Error (

R M S E

) between actual and predicted values, as well as Directional Accuracy (

D A

). We show that a single time series from a dataset can be used to adequately train the networks if the sequences in the dataset contain patterns that repeat, even with certain variation, and are properly processed. For 1-step ahead and 20-step ahead forecasts, LSTM and GRU networks significantly outperform a baseline on the Activities dataset. The baseline simply repeats the last available value. On the stock market dataset, the networks perform just as the baseline, possibly due to the nature of these series. We release the datasets used as well as the implementation with all experiments performed to enable future comparisons and to make our research reproducible. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 286 KiB

Open AccessProceeding Paper

Outliers Impact on Parameter Estimation of Gaussian and Non-Gaussian State Space Models: A Simulation Study

by Fernanda Catarina Pereira, Arminda Manuela Gonçalves and Marco Costa

Eng. Proc. 2022, 18(1), 31; https://doi.org/10.3390/engproc2022018031 - 22 Jun 2022

Viewed by 1041

Abstract

State space models are powerful and quite flexible tools that allow systems that vary significantly over time due to their formulation to be dealt with, because the models’ parameters vary over time. Assuming a known distribution of errors, in particular the Gaussian distribution, [...] Read more.

State space models are powerful and quite flexible tools that allow systems that vary significantly over time due to their formulation to be dealt with, because the models’ parameters vary over time. Assuming a known distribution of errors, in particular the Gaussian distribution, parameter estimation is usually performed by maximum likelihood. However, in time series data, it is common to have discrepant values that can impact statistical data analysis. This paper presents a simulation study with several scenarios to find out in which situations outliers can affect the maximum likelihood estimators. The results obtained were evaluated in terms of the difference between the maximum likelihood estimate and the true value of the parameter and the rate of valid estimates. It was found that both for Gaussian and exponential errors, outliers had more impact in two situations: when the sample size is small and the autoregressive parameter is close to 1, and when the sample size is large and the autoregressive parameter is close to 0.25. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

7 pages, 328 KiB

Open AccessProceeding Paper

Time Series Sampling

by Florian Combes, Ricardo Fraiman and Badih Ghattas

Eng. Proc. 2022, 18(1), 32; https://doi.org/10.3390/engproc2022018032 - 23 Jun 2022

Cited by 1 | Viewed by 2334

Abstract

Some complex models are frequently employed to describe physical and mechanical phenomena. In this setting, we have an input X, which is a time series, and an output

Y = f (X)

where f is a very complicated function, whose [...] Read more.

Some complex models are frequently employed to describe physical and mechanical phenomena. In this setting, we have an input X, which is a time series, and an output

Y = f (X)

where f is a very complicated function, whose computational cost for every new input is very high. We are given two sets of observations of X,

S_{1}

and

S_{2}

of different sizes such that only

f (S_{1})

is available. We tackle the problem of selecting a subsample

S_{3} \in S_{2}

of a smaller size on which to run the complex model f and such that distribution of

f (S_{3})

is close to that of

f (S_{1})

. We adapt to this new framework five algorithms introduced in a previous work "Subsampling under Distributional Constraints" to solve this problem and show their efficiency using time series data. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 1148 KiB

Open AccessProceeding Paper

Modelling a Continuous Time Series with FOU(p) Processes

by Juan Kalemkerian

Eng. Proc. 2022, 18(1), 33; https://doi.org/10.3390/engproc2022018033 - 23 Jun 2022

Viewed by 806

Abstract

In this work we summarize the knowledge about FOU(p) processes (fractional iterated Ornstein–Uhlenbeck processes of order emphp). Fractional Ornstein–Uhlenbeck processes are a particular case of FOU(p) processes (when

p = 1

). FOU(p) processes are able to model time series [...] Read more.

In this work we summarize the knowledge about FOU(p) processes (fractional iterated Ornstein–Uhlenbeck processes of order emphp). Fractional Ornstein–Uhlenbeck processes are a particular case of FOU(p) processes (when

p = 1

). FOU(p) processes are able to model time series with both long- and short-range dependence. We give the definition, the main theoretical properties, and a procedure for estimating the parameters consistently. We also show how to model a continuous time series with FOU(p) processes, and we give an example of an application. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 2825 KiB

Open AccessProceeding Paper

PV Energy Prediction in 24 h Horizon Using Modular Models Based on Polynomial Conversion of the L-Transform PDE Derivatives in Node-by-Node-Evolved Binary-Tree Networks

by Ladislav Zjavka and Václav Snášel

Eng. Proc. 2022, 18(1), 34; https://doi.org/10.3390/engproc2022018034 - 27 Jun 2022

Cited by 3 | Viewed by 994

Abstract

Accurate daily photovoltaic (PV) power predictions are challenging as near-ground atmospheric processes include complicated chaotic interactions among local factors (ground temperature, cloudiness structure, humidity, visibility factor, etc.). Fluctuations in solar irradiance resulting from the cloud structure dynamics are influenced by many uncertain parameters, [...] Read more.

Accurate daily photovoltaic (PV) power predictions are challenging as near-ground atmospheric processes include complicated chaotic interactions among local factors (ground temperature, cloudiness structure, humidity, visibility factor, etc.). Fluctuations in solar irradiance resulting from the cloud structure dynamics are influenced by many uncertain parameters, which can be described by differential equations. Recent artificial intelligence (AI) computational tools allow us to transform and post-validate forecast data from numerical weather prediction (NWP) systems to estimate PV power generation in relation to on-site local specifics. However, local NWP models are usually produced each six hours to simulate the progress of main weather quantities in a medium-scale target area. Their delay usually covers several hours, further increasing the inadequate operational quality required in PV plants. All-day prediction models perform better, if they are developed with the last historical weather and PV data. Differential polynomial neural network (D-PNN) is a recently designed computational method, based on a new learning approach, which allows us to represent complicated data relations contained in local weather patterns to account for irregular phenomena. D-PNN combines two-input variables to split the partial differential equation (PDE), defined in the general order k and n variables, into partition elements of two-input node PDEs of recognized order and type. The node-determined sub-PDEs can be easily converted using operator calculus (OC), in several types of predefined convert schemes, to define unknown node functions expressed in the Laplace images form Application of the inverse L-transformation formula to the L-converts results in obtaining the prime function originals. D-PNN elicits a progressive modular tree structure to assess one-by-one the optimal PDE node solutions to be inserted in the sum output of the overall expanded computing model. Statistical modular models are the result of learning schemes of preadjusted day data records from various observational localities. They are applied after testing to the rest of unseen daily series of known data to compute estimations of clear-sky index (CSI) in the 24 h input-delayed time-sequences. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 304 KiB

Open AccessProceeding Paper

Modelling the Number of Daily Stock Transactions Using a Novel Time Series Model

by Sunecher Yuvraj

Eng. Proc. 2022, 18(1), 35; https://doi.org/10.3390/engproc2022018035 - 27 Jun 2022

Viewed by 895

Abstract

This paper focusses on the impact of the COVID-19 on the Stock Exchange of Mauritius (SEM) by modelling the number of daily stock transactions of two banks. Hence, a non-stationary bivariate integer-valued autoregressive and moving average of order 1 (BINARMA (1,1)) process with [...] Read more.

This paper focusses on the impact of the COVID-19 on the Stock Exchange of Mauritius (SEM) by modelling the number of daily stock transactions of two banks. Hence, a non-stationary bivariate integer-valued autoregressive and moving average of order 1 (BINARMA (1,1)) process with COM-Poisson (CMP) innovations (BINARMA (1,1) CMP) is introduced. The conditional maximum likelihood (CML) approach is used to estimate the model parameters. The novel model is applied on the intra-day trading of two banking stocks. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

15 pages, 3354 KiB

Open AccessProceeding Paper

Improving the Predictive Power of Historical Consistent Neural Networks

by Rockefeller Rockefeller, Bubacarr Bah, Vukosi Marivate and Hans-Georg Zimmermann

Eng. Proc. 2022, 18(1), 36; https://doi.org/10.3390/engproc2022018036 - 27 Jun 2022

Cited by 1 | Viewed by 1494

Abstract

The Historical Consistent Neural Networks (HCNN) are an extension of the standard Recurrent Neural Networks (RNN): they allow the modeling of highly-interacting dynamical systems across multiple time scales. HCNN do not draw any distinction between inputs and outputs, but model observables embedded in [...] Read more.

The Historical Consistent Neural Networks (HCNN) are an extension of the standard Recurrent Neural Networks (RNN): they allow the modeling of highly-interacting dynamical systems across multiple time scales. HCNN do not draw any distinction between inputs and outputs, but model observables embedded in the dynamics of a large state space. In this paper, we propose to improve the predictive power of the (Vanilla) HCNN using three methods: (1) HCNN with Partial Teacher Forcing, (2) HCNN with Sparse State Transition Matrix, and (3) a Long Short Term Memory Formulation of HCNN. We investigated the effect of those long memory improvement methods on three chaotic time-series mathematically generated from the Rabinovich–Fabrikant, the Rossler System and the Lorenz system. To complement our study, we compared the accuracy of the different HCNN variants with well-known recurrent neural networks methods such as Vanilla RNN and LSTM for the same prediction tasks. Overall, our results show that the Vanilla HCNN is superior to RNN and LSTM. This is even more the case if you include the above long memory extensions (1), (2) and (3). We demonstrate that (1) and (3) are superior for the modeling of our chaotic dynamical systems. We show that for these deterministic systems, the ensembles are narrowed. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

12 pages, 388 KiB

Open AccessProceeding Paper

Exploration of Different Time Series Models for Soccer Athlete Performance Prediction

by Siarhei Kulakou, Nourhan Ragab, Cise Midoglu, Matthias Boeker, Dag Johansen, Michael A. Riegler and Pål Halvorsen

Eng. Proc. 2022, 18(1), 37; https://doi.org/10.3390/engproc2022018037 - 29 Jun 2022

Cited by 4 | Viewed by 1761

Abstract

Professional sports achievements combine not only the individual physical abilities of athletes but also many modern technologies in areas such as medicine, equipment production, nutrition, and physical and mental health monitoring. In this work, we address the problem of predicting soccer players’ ability [...] Read more.

Professional sports achievements combine not only the individual physical abilities of athletes but also many modern technologies in areas such as medicine, equipment production, nutrition, and physical and mental health monitoring. In this work, we address the problem of predicting soccer players’ ability to perform, from subjective self-reported wellness parameters collected using a commercially deployed digital health monitoring system called PmSys. We use 2 years of data from two Norwegian female soccer teams, where players have reported daily ratings for their readiness-to-play, mood, stress, general muscle soreness, fatigue, sleep quality, and sleep duration. We explore various time series models with the goal of predicting readiness, employing both a univariate approach and a multivariate approach. We provide an experimental comparison of different time series models, such as purely recurrent models, models of mixed recursive convolutional types, ensemble of deep CNN models, and multivariate versions of the recurrent models, in terms of prediction performance, with a focus on detecting peaks. We use different input and prediction windows to compare the accuracy of next-day predictions and next-week predictions. We also investigate the potential of using models built on data from the whole team for making predictions about individual players, as compared to using models built on the data from the individual player only. We tackle the missing data problem by various methods, including the replacement of all gaps with zeros, filling in repeated values, as well as removing all gaps and concatenating arrays. Our case study on athlete monitoring shows that a number of time series analysis models are able to predict readiness with high accuracy in near real-time. Equipped with such insight, coaches and trainers can better plan individual and team training sessions, and perhaps avoid over training and injuries. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 435 KiB

Open AccessProceeding Paper

The Bootstrap for Testing the Equality of Two Multivariate Stochastic Processes with an Application to Financial Markets

by Ángel López-Oriona and José A. Vilar

Eng. Proc. 2022, 18(1), 38; https://doi.org/10.3390/engproc2022018038 - 8 Jul 2022

Viewed by 752

Abstract

The problem of testing the equality of generating processes of two multivariate time series is addressed in this work. To this end, we construct two tests based on a distance measure between stochastic processes. The metric is defined in terms of the quantile [...] Read more.

The problem of testing the equality of generating processes of two multivariate time series is addressed in this work. To this end, we construct two tests based on a distance measure between stochastic processes. The metric is defined in terms of the quantile cross-spectral densities of both processes. A proper estimate of this dissimilarity is the cornerstone of the proposed tests. Both techniques are based on the bootstrap. Specifically, extensions of the moving block bootstrap and the stationary bootstrap are used for their construction. The approaches are assessed in a broad range of scenarios under the null and the alternative hypotheses. The results from the analyses show that the procedure based on the stationary bootstrap exhibits the best overall performance in terms of both size and power. The proposed techniques are used to answer the question regarding whether or not the dotcom bubble crash of the 2000s permanently impacted global market behavior. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

13 pages, 2583 KiB

Open AccessProceeding Paper

Using Forecasting Methods on Crime Data: The SKALA Approach of the State Office for Criminal Investigation of North Rhine-Westphalia

by Kai Seidensticker and Katharina Schwarz

Eng. Proc. 2022, 18(1), 39; https://doi.org/10.3390/engproc2022018039 - 11 Jul 2022

Cited by 1 | Viewed by 1883

Abstract

In this article, we introduce the topic of crime forecasting performed in North Rhine-Westphalia, Germany. We give a brief overview of three forecasting methods used in theory and practice: predictive policing, risk terrain modeling, and time series analysis. As a result, spatio-temporally-based statistical [...] Read more.

In this article, we introduce the topic of crime forecasting performed in North Rhine-Westphalia, Germany. We give a brief overview of three forecasting methods used in theory and practice: predictive policing, risk terrain modeling, and time series analysis. As a result, spatio-temporally-based statistical techniques offered high potential to optimize operational and strategic planning for policing. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

13 pages, 1080 KiB

Open AccessProceeding Paper

Reconstructed Phase Spaces and LSTM Neural Network Ensemble Predictions

by Sebastian Raubitzek and Thomas Neubauer

Eng. Proc. 2022, 18(1), 40; https://doi.org/10.3390/engproc2022018040 - 25 Jul 2022

Viewed by 980

Abstract

We present a novel approach that combines the concept of reconstructed phase spaces with neural network time-series predictions. The presented methodology aims to reduce the parametrization problem of neural networks and improve autoregressive neural network time-series predictions. First, the idea is to interpolate [...] Read more.

We present a novel approach that combines the concept of reconstructed phase spaces with neural network time-series predictions. The presented methodology aims to reduce the parametrization problem of neural networks and improve autoregressive neural network time-series predictions. First, the idea is to interpolate a dataset based on its reconstructed phase space properties and then filter an ensemble prediction based on its phase space properties. The corresponding ensemble predictions are made using randomly parameterized LSTM (Long Short-Term Memory) neural networks. These neural networks then produce a multitude of auto-regressive predictions, which are then filtered to achieve a smooth reconstructed phase space trajectory. Thus, we can circumvent the problem of parameterizing the neural network for each dataset individually. Here, the interpolation and the ensemble prediction aim to produce a smooth trajectory in a reconstructed phase space. The best results are compared to a single hidden layer LSTM neural network and benchmark results from the literature. The results show that the baseline predictions are outperformed for all three discussed datasets, and one of the benchmark results from the literature is bested by the presented approach. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

17 pages, 1098 KiB

Open AccessProceeding Paper

Dynamic Asymmetric Causality Tests with an Application

by Abdulnasser Hatemi-J

Eng. Proc. 2022, 18(1), 41; https://doi.org/10.3390/engproc2022018041 - 2 Aug 2022

Cited by 8 | Viewed by 1795

Abstract

Testing for causation—defined as the preceding impact of the past value(s) of one variable on the current value of another when all other pertinent information is accounted for—is increasingly utilized in empirical research using the time-series data in different scientific disciplines. A relatively [...] Read more.

Testing for causation—defined as the preceding impact of the past value(s) of one variable on the current value of another when all other pertinent information is accounted for—is increasingly utilized in empirical research using the time-series data in different scientific disciplines. A relatively recent extension of this approach has been allowing for potential asymmetric impacts, since this is harmonious with the way reality operates, in many cases. The current paper maintains that it is also important to account for the potential change in the parameters when asymmetric causation tests are conducted, as several reasons exist for changing the potential causal connection between the variables across time. Therefore, the current paper extends the static asymmetric causality tests by making them dynamic via the use of subsamples. An application is also provided consistent with measurable definitions of economic, or financial bad, as well as good, news and their potential causal interactions across time. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

4 pages, 261 KiB

Open AccessProceeding Paper

Coarse Grain Spectral Analysis for the Low-Amplitude Signature of Multiperiodic Stellar Pulsators

by Sebastià Barceló Forteza, Javier Pascual-Granado, Juan Carlos Suárez, Antonio García Hernández and Mariel Lares-Martiz

Eng. Proc. 2022, 18(1), 42; https://doi.org/10.3390/engproc2022018042 - 22 Aug 2022

Viewed by 866

Abstract

Coarse Grain Spectral Analysis (CGSA) can explain the possible multiscaling nature of the thousands of low-amplitude peaks observed in the power spectra of some pulsating stars. Space-based observations allowed for the scientific community to find this kind of structure thanks to their long-duration, [...] Read more.

Coarse Grain Spectral Analysis (CGSA) can explain the possible multiscaling nature of the thousands of low-amplitude peaks observed in the power spectra of some pulsating stars. Space-based observations allowed for the scientific community to find this kind of structure thanks to their long-duration, high-photometric precision and duty cycle compared to observations from the ground. Although these time series are far from perfect (outliers, trends, gaps, etc.), we used our own data preprocessing method, known as the 2K+1 stage interpolation method, to improve the background noise up to a factor 14, avoiding spurious effects. We applied both techniques, the 2K+1 stage method and the CGSA analysis, to shed some light on a real problem regarding stellar seismology: finding the physical nature of the low-amplitude signature for multiperiodic stellar pulsators. Full article

(This article belongs to the Proceedings of The 8th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Eng. Proc., 2022, ITISE 2022

The 8th International Conference on Time Series and Forecasting

Editorial

Other

Further Information

Guidelines

MDPI Initiatives

Follow MDPI