Search

In this paper, a novel combination of deep learning recurrent neural network and Lyapunov time is proposed to forecast the consumption of electricity load, in Greece, in normal/abrupt change value areas. Our method verifies the chaotic behavior of load time series through chaos time series analysis and with the application of deep learning recurrent neural networks produces predictions for 10 and 20 days ahead. Specifically, four different neural network models constructed (a) feed forward neural network, (b) gated recurrent unit (GRU) neural network, (c) long short-term memory (LSTM) recurrent and (d) bidirectional LSTM neural network to implement the prediction in a prediction horizon, produced through the extraction of maximum Lyapunov exponent. We constructed sequences of algorithms to feed the neural networks, creating three scenarios (a) 1-step, (b) 10-step and (c) 20-step sequences. For each neural network model, we used its predictions as inputs to predict steps forward, iteratively, to examine the accuracy of the proposed models, for horizons that are both inside and outside to that defined by Lyapunov time. The results show that the deep learning GRU neural network produces iterative predictions of high accuracy and stability, following the trend evolution of actual values, even outside the safe horizon for 1-step and 10-step cases. [ABSTRACT FROM AUTHOR]

Several roundabout capacity methods and approaches have been proposed until now. They are mainly based either on regression equations of observed capacity and gap acceptance or on stochastic models through simulation techniques. However, all of them rely on different variables and factors and are being adapted to local driving behavior. Hence, it is not clear if existing techniques are appropriate for reliable capacity estimations and optimal design of Greek roundabouts. This paper presents the results of an experimental research that has been conducted as a first step in the optimization of roundabouts capacity estimation, based on a dataset of 11 roundabouts in Greece. The study firstly aims to understand what geometric roundabout features and driving behavior parameters influence capacity. Artificial neural networks (NN) were tested and developed to predict accurate roundabout capacities. Appropriate roundabouts were selected for the analysis and their operational performance was filmed using a UAV and a stabilized camera during peak periods. Video image processing techniques and algorithms allowed the extraction of empirical data (traffic flows, gap acceptance parameters) and accurate geometric characteristics. The results demonstrate that artificial neural networks can predict the capacity of roundabouts accurately. [ABSTRACT FROM AUTHOR]

This paper presents and assesses various machine learning methods that aim at predicting path loss in rural environment. For this purpose, models such as artificial neural network (ANN), support vector regression (SVR), random forest (RF), and bagging with k-nearest neighbor (B-kNN) learners, are exploited and evaluated. They are trained and tested with path loss data collected from an extensive measurement campaign that have been carried out in diverse rural areas in Greece. The results demonstrate that all the proposed machine learning models outperform the empirical ones, exhibiting, in any case, root-mean-square-error (RMSE) values between 4.0 and 6.5 dB. The poorest prediction of the measured data is encountered for SVR with Polynomial kernel. Furthermore, B-kNN and RF algorithms preserve comparable path loss approximations with remarkably low RMSE on the order of 4.2–4.3 dB. The error metrics also reveal that increasing the number of hidden layers in ANNs, their performance is gradually enhanced. However, deeper layouts with more than three hidden layers do not markedly improve any further the prediction accuracy. Finally, the best prediction is achieved when employing a three-hidden layered ANN with 51 neurons evenly distributed among the layers. The specific layout exhibits the lowest RMSE value (4.0 dB), thus being highly recommended for accurate path loss predictions in rural locations. [ABSTRACT FROM AUTHOR]

This paper looks into the Greek-Turkish arms race a decade after an earlier contribution to the issue that relied heavily on artificial neural networks. The time period between the two papers contributes to the reliability of the results derived, not just by increasing the number of observations, but also mainly by incorporating the progress made in the realm of artificial intelligence. The focus on the case of both countries unlike the paper mentioned above that dealt with just the Greek side provides ample room for comparative purposes regarding the determinants of defense expenditure on both sides. The results derived in terms of input significance estimation support the findings of the earlier research as indicated above, pointing to the leading role of the demographic preponderance of Turkey over Greece. The paper also points to the fact that 10 years later, Turkey continues to set the arms race rules against its rival by determining the defense expenditure of Greece, whereas the role of the latter in affecting the military spending of Turkey is non-existent. [ABSTRACT FROM AUTHOR]

This paper introduced a new ensemble learning approach, based on evolutionary fuzzy cognitive maps (FCMs), artificial neural networks (ANNs), and their hybrid structure (FCM-ANN), for time series prediction. The main aim of time series forecasting is to obtain reasonably accurate forecasts of future data from analyzing records of data. In the paper, we proposed an ensemble-based forecast combination methodology as an alternative approach to forecasting methods for time series prediction. The ensemble learning technique combines various learning algorithms, including SOGA (structure optimization genetic algorithm)-based FCMs, RCGA (real coded genetic algorithm)-based FCMs, efficient and adaptive ANNs architectures, and a hybrid structure of FCM-ANN, recently proposed for time series forecasting. All ensemble algorithms execute according to the one-step prediction regime. The particular forecast combination approach was specifically selected due to the advanced features of each ensemble component, where the findings of this work evinced the effectiveness of this approach, in terms of prediction accuracy, when compared against other well-known, independent forecasting approaches, such as ANNs or FCMs, and the long short-term memory (LSTM) algorithm as well. The suggested ensemble learning approach was applied to three distribution points that compose the natural gas grid of a Greek region. For the evaluation of the proposed approach, a real-time series dataset for natural gas prediction was used. We also provided a detailed discussion on the performance of the individual predictors, the ensemble predictors, and their combination through two well-known ensemble methods (the average and the error-based) that are characterized in the literature as particularly accurate and effective. The prediction results showed the efficacy of the proposed ensemble learning approach, and the comparative analysis demonstrated enough evidence that the approach could be used effectively to conduct forecasting based on multivariate time series. [ABSTRACT FROM AUTHOR]

This paper aimed to investigate the potential of using spectral and textural features in the Meteosat Second Generation-Spinning Enhanced Visible and Infrared Imager (MSG-SEVIRI) data for developing techniques capable of classifying convective and stratiform rain areas in the satellite image. Two different classification methods were introduced that use the brightness temperature (BT) T and brightness temperature differences T- T T- T T- T T- T T- T T- T, and T- T as spectral parameters, along with textural parameters derived from the thermal infrared MSG-SEVIRI channel to discriminate between convective and stratiform rainy clouds. The first is an algorithm based on the probability of convective rainfall (PCR) for each pixel of the MSG satellite data and the second is an artificial neural network multilayer perceptron (MLP) model that relies on the correlation of satellite data with convective and stratiform rain. Both schemes were trained using as reference convective/stratiform classification data from 88 stations in Greece for 20 rainy days with high convective activity and evaluated against an independent dataset of gauge data for six rainy days. Two PCR and two MLP algorithms were constructed based on the previous results: the PCR1 and MLP1 algorithms that use only spectral measures and the PCR2 and MLP2 algorithms based on both spectral and textural measures. It was found that the introduction of textural parameters as additional information to a technique (PCR2 and MLP2 ) works in improving the delineation between convective and stratiform rainy pixels compared to the techniques based only on spectral information (PCR1 and MLP1). The comparison of the two schemes revealed a clear outperformance of the MLP techniques over the PCR techniques. Best skill is provided by MLP2 (POD = 74.5 %, FAR = 44.3 %, POFD = 22.5 %, CSI = 0.47, ETS = 0.31, bias = 1.34) followed by MLP1 and the two PCR techniques. All algorithms overestimate the convective rain occurrences observed by the rain gauge network. These findings showed that the combined use of spectral and textural features in the MSG-SEVIRI imagery can be beneficial for the classification of convective and stratiform precipitating clouds. [ABSTRACT FROM AUTHOR]

Abstract: This paper presents an artificial intelligence approach integrated with geographical information systems (GISs) for modeling urban evolution. Fuzzy logic and neural networks are used to provide a synthetic spatiotemporal methodology for the analysis, prediction and interpretation of urban growth. The proposed urban model takes into account the changes over time in population and building use patterns. A GIS is used for handling the spatial and temporal data, performing contingency analysis and mapping the results. Spatial entities with similar characteristics are grouped together in clusters by the use of a fuzzy c-means algorithm. Each cluster represents a specific level of urban growth and development. A two-layer feed-forward multilayer perceptron artificial neural network is then used to predict urban growth. The model, applied to the prefecture of Attica, Greece, delineates the current and future evolution trends of the Athens metropolitan area, which are illustrated by maps of the urban growth dynamics. The proposed methodology aims to assist planners and decision makers in gaining insight into the transition from rural to urban. [Copyright &y& Elsevier]

Abstract: In this paper we propose a methodology consisting of specific computational intelligence methods, i.e. principal component analysis and artificial neural networks, in order to inter-compare air quality and meteorological data, and to forecast the concentration levels for environmental parameters of interest (air pollutants). We demonstrate these methods to data monitored in the urban areas of Thessaloniki and Helsinki in Greece and Finland, respectively. For this purpose, we applied the principal component analysis method in order to inter-compare the patterns of air pollution in the two selected cities. Then, we proceeded with the development of air quality forecasting models for both studied areas. On this basis, we formulated and employed a novel hybrid scheme in the selection process of input variables for the forecasting models, involving a combination of linear regression and artificial neural networks (multi-layer perceptron) models. The latter ones were used for the forecasting of the daily mean concentrations of PM10 and PM2.5 for the next day. Results demonstrated an index of agreement between measured and modelled daily averaged PM10 concentrations, between 0.80 and 0.85, while the kappa index for the forecasting of the daily averaged PM10 concentrations reached 60% for both cities. Compared with previous corresponding studies, these statistical parameters indicate an improved performance of air quality parameters forecasting. It was also found that the performance of the models for the forecasting of the daily mean concentrations of PM10 was not substantially different for both cities, despite the major differences of the two urban environments under consideration. [Copyright &y& Elsevier]

Abstract: Missing data are omnipresent in forestry research, and this poses problems in the analysis of primary data. Many statistical problems have been viewed as missing data problems. To cope with incomplete data, several methods are currently being used. They are all based on assumptions some of which might not be valid in a particular case. The choice mostly depends on the objective of the study. Considerable mensuration research is motivated by the need for yield projections that can support forest management decisions. This paper is focused on a new approach for filling gaps in diameter measurements on standing tree boles. Dealing with this problem, an attempt was made to examine the applicability of artificial neural network models for missing data estimation and to use the estimated values in the subsequent analysis. The procedure that should be followed in the development of such models is outlined. The results show good performance of the examined ANN models compared to regression treatments for missing data and ANN models demonstrate their adequacy and potential for filling gaps in diameter measurements on standing tree boles. The ANN models applied in this study are sufficiently general and have great potential to be applicable for estimating the missing values of many variables in environmental applications. [Copyright &y& Elsevier]

Abstract: Air pollution modelling is necessary for simulating the atmospheric environment system in terms of pollutants and meteorological conditions, taking into account the nonlinearities of the underlying phenomena. In the current paper, Artificial Neural Networks are used for modelling ozone, and for simulating its behaviour in relation to other atmospheric parameters of interest, for the city of Thessaloniki, Greece. This behaviour is also investigated with the aid of Principal Component Analysis (PCA). Results suggest the operational capabilities of such models, and the research potential in the application of computational intelligence methods for the environmental sector. [Copyright &y& Elsevier]

Modeling the elasticity modulus of unbound granular pavement materials has attracted significant research interest because of its importance in pavement design particularly in PPP/BOT projects. These efforts have been hampered by three factors: (i) inability to capture the correlations between the asphalt and granular layers and the subgrade, (ii) inadequate modeling of the effects of external factors on the elasticity modulus of unbound materials, and (iii) widespread use of linear statistical relationships to model a complex and non-linear phenomenon. In this paper genetically optimized neural networks and falling weight deflectometer (FWD) back-analysis results from a newly constructed BOT project in Athens, Greece, are employed in order to evaluate pavement section design parameters. It is shown that parameter values adopted during design do not co-inside with those observed from the back-analysis studies. Further, the results indicate that the relative estimation error for the modulus of elasticity of the unbound material does not exceed 25%, while the correlation between actual and predicted values is 86%, both suggesting that the proposed approach models the physical phenomenon adequately, a finding with important practical implications particularly in PPP projects. [ABSTRACT FROM AUTHOR]

• A model for estimating Life Cycle Cost of road tunnels. • A homogeneous database of 32 road tunnels, with a total length of 55 km. • Artificial Neural Networks are employed to estimate final costs. • Rapid cost pre-estimation of alternatives at early stages. • Development of tunnel operation & maintenance strategy. The estimation of road tunnel costs is crucial for motorway operators and transport infrastructure project managers. This paper presents a life cycle model for estimating costs in road tunnels. The research is based on a data set of 32 dual bore road tunnels, with a total length of 55 km, which are part of the Egnatia Odos Motorway in Northern Greece. The factors affecting civil engineering construction works and electromechanical installations are analyzed in great detail. Energy consumption and operation and maintenance costs are examined. Additionally, replacement costs are studied for an operation period of 75 years. Overall, Artificial Neural Networks are employed to estimate final costs in relation to tunnel design input parameters. The life cycle cost model uses depreciation and price escalation on all costs. It is concluded that the proposed framework can assist tunnel operators and asset managers in reliable costing of alternatives as well as developing operation and maintenance strategies. [ABSTRACT FROM AUTHOR]

The aim of this paper is the generation of a time-series based statistical data-driven procedure in order to track an outbreak. At first are used univariate time series models in order to predict the evolution of the reported cases. Moreover, are considered combinations of the models in order to provide more accurate and robust results. Additionally, statistical probability distributions are considered in order to generate future scenarios. Final step is the build and use of an epidemiological model (tSIR) and the calculation of an epidemiological ratio (R 0) for estimating the termination of the outbreak. The time series models include Exponential Smoothing and ARIMA approaches from the classical models, also Feed-Forward Artificial Neural Networks and Multivariate Adaptive Regression Splines from the machine learning toolbox. Combinations include simple mean, Newbolt-Granger and Bates-Granger approaches. Finally, the tSIR model and the R 0 ratio are used for estimating the spread and the reversion of the pandemic. The suggested procedure is used to track the COVID-19 epidemic in Greece. This epidemic has appeared in China in December 2019 and has been widespread since then to all over the world. Greece is the center of this empirical study as is considered an early successful paradigm of resistance against the virus. [ABSTRACT FROM AUTHOR]

Lake Koronia is located in the North part of Greece and is protected by the Ramsar Convention of wetlands. A deficit in the water balance has been presented at the last twenty years due to the excessive water consumption for agricultural uses. This research is an attempt to simulate water flow with MIKE SHE model in order to observe how the water is allocated in the study area. The results of water flow module used for the estimation of Lake's water balance for 4 hydrological years (2008-2012). Furthermore the Artificial Neural Networks (ANNs) was used for the prediction of water flow in two sub-catchments. The coefficient correlation (R) was found for Bogdanas (0.9) and Kolxikos (0.86). The Root Mean Square Error (RMSE) and the Mean Absolute Percentages Error (MAPE) were also calculated in order to evaluate the quality of the ANNs results. [ABSTRACT FROM AUTHOR]