scholarly journals Methods of Constructing Time Series for Predicting Local Time Scales by Means of a GMDH-Type Neural Network

2021 ◽  
Vol 11 (12) ◽  
pp. 5615
Author(s):  
Łukasz Sobolewski ◽  
Wiesław Miczulski
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Local Time ◽  
Time Scales ◽  
Time Scale ◽  
Prediction Accuracy ◽  
Prediction Horizon

Ensuring the best possible stability of UTC(k) (local time scale) and its compliance with the UTC scale (Universal Coordinated Time) forces predicting the [UTC-UTC(k)] deviations, the article presents the results of work on two methods of constructing time series (TS) for a neural network (NN), increasing the accuracy of UTC(k) prediction. In the first method, two prepared TSs are based on the deviations determined according to the UTC scale with a 5-day interval. In order to improve the accuracy of predicting the deviations, the PCHIP interpolating function is used in subsequent TSs, obtaining TS elements with a 1-day interval. A limitation in the improvement of prediction accuracy for these TS has been a too large prediction horizon. The introduction in 2012 of the additional UTC Rapid scale by BIPM makes it possible to shorten the prediction horizon, and the building of two TSs has been proposed according to the second method. Each of them consists of two subsets. The first subset is based on deviations determined according to the UTC scale, the second on the UTC Rapid scale. The research of the proposed TS in the field of predicting deviations for the Polish Timescale by means of GMDH-type NN shows that the best accuracy of predicting the deviations has been achieved for TS built according to the second method.

scholarly journals Locally and Remotely Forced Subtropical AMOC Variability: A Matter of Time Scales

2020 ◽  
Vol 33 (12) ◽  
pp. 5155-5172
Author(s):  
Quentin Jamet ◽  
William K. Dewar ◽  
Nicolas Wienders ◽  
Bruno Deremble ◽  
Sally Close ◽  
...  
Keyword(s):  
Time Series ◽  
Time Scales ◽  
North Atlantic ◽  
Time Scale ◽  
Low Frequency ◽  
Meridional Overturning Circulation ◽  
Open Boundary ◽  
Overturning Circulation ◽  
Decadal Scale ◽  
Meridional Overturning

AbstractMechanisms driving the North Atlantic meridional overturning circulation (AMOC) variability at low frequency are of central interest for accurate climate predictions. Although the subpolar gyre region has been identified as a preferred place for generating climate time-scale signals, their southward propagation remains under consideration, complicating the interpretation of the observed time series provided by the Rapid Climate Change–Meridional Overturning Circulation and Heatflux Array–Western Boundary Time Series (RAPID–MOCHA–WBTS) program. In this study, we aim at disentangling the respective contribution of the local atmospheric forcing from signals of remote origin for the subtropical low-frequency AMOC variability. We analyze for this a set of four ensembles of a regional (20°S–55°N), eddy-resolving (1/12°) North Atlantic oceanic configuration, where surface forcing and open boundary conditions are alternatively permuted from fully varying (realistic) to yearly repeating signals. Their analysis reveals the predominance of local, atmospherically forced signal at interannual time scales (2–10 years), whereas signals imposed by the boundaries are responsible for the decadal (10–30 years) part of the spectrum. Due to this marked time-scale separation, we show that, although the intergyre region exhibits peculiarities, most of the subtropical AMOC variability can be understood as a linear superposition of these two signals. Finally, we find that the decadal-scale, boundary-forced AMOC variability has both northern and southern origins, although the former dominates over the latter, including at the site of the RAPID array (26.5°N).

scholarly journals Sub-daily runoff simulations with parameters inferred at the daily time scale

2015 ◽  
Vol 12 (8) ◽  
pp. 7437-7467 ◽  
Author(s):  
J. E. Reynolds ◽  
S. Halldin ◽  
C. Y. Xu ◽  
J. Seibert ◽  
A. Kauffeldt
Keyword(s):  
Time Series ◽  
Numerical Method ◽  
Time Scales ◽  
Time Scale ◽  
Input Data ◽  
Flood Forecasting ◽  
Euler Method ◽  
Time Stepping ◽  
Parameter Values ◽  
Daily Time

Abstract. Concentration times in small and medium-sized watersheds (~ 100–1000 km2) are commonly less than 24 h. Flood-forecasting models then require data at sub-daily time scales, but time-series of input and runoff data with sufficient lengths are often only available at the daily time scale, especially in developing countries. This has led to a search for time-scale relationships to infer parameter values at the time scales where they are needed from the time scales where they are available. In this study, time-scale dependencies in the HBV-light conceptual hydrological model were assessed within the generalized likelihood uncertainty estimation (GLUE) approach. It was hypothesised that the existence of such dependencies is a result of the numerical method or time-stepping scheme used in the models rather than a real time-scale-data dependence. Parameter values inferred showed a clear dependence on time scale when the explicit Euler method was used for modelling at the same time steps as the time scale of the input data (1–24 h). However, the dependence almost fully disappeared when the explicit Euler method was used for modelling in 1 h time steps internally irrespectively of the time scale of the input data. In other words, it was found that when an adequate time-stepping scheme was implemented, parameter sets inferred at one time scale (e.g., daily) could be used directly for runoff simulations at other time scales (e.g., 3 or 6 h) without any time scaling and this approach only resulted in a small (if any) model performance decrease, in terms of Nash–Sutcliffe and volume-error efficiencies. The overall results of this study indicated that as soon as sub-daily driving data can be secured, flood forecasting in watersheds with sub-daily concentration times is possible with model-parameter values inferred from long time series of daily data, as long as an appropriate numerical method is used.

scholarly journals A Dual-Attention Recurrent Neural Network Method for Deep Cone Thickener Underflow Concentration Prediction

Sensors ◽  
2020 ◽  
Vol 20 (5) ◽  
pp. 1260
Author(s):  
Zhaolin Yuan ◽  
Jinlong Hu ◽  
Di Wu ◽  
Xiaojuan Ban
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Prediction Accuracy ◽  
Time Series Prediction ◽  
Experimental Result ◽  
Common Error ◽  
Neural Network Method ◽  
Multiple Sensors ◽  
Network Method ◽  
Deep Cone Thickener

This paper focuses on the time series prediction problem for underflow concentration of deep cone thickener. It is commonly used in the industrial sedimentation process. In this paper, we introduce a dual attention neural network method to model both spatial and temporal features of the data collected from multiple sensors in the thickener to predict underflow concentration. The concentration is the key factor for future mining process. This model includes encoder and decoder. Their function is to capture spatial and temporal importance separately from input data, and output more accurate prediction. We also consider the domain knowledge in modeling process. Several supplementary constructed features are examined to enhance the final prediction accuracy in addition to the raw data from sensors. To test the feasibility and efficiency of this method, we select an industrial case based on Industrial Internet of Things (IIoT). This Tailings Thickener is from FLSmidth with multiple sensors. The comparative results support this method has favorable prediction accuracy, which is more than 10% lower than other time series prediction models in some common error indices. We also try to interpret our method with additional ablation experiments for different features and attention mechanisms. By employing mean absolute error index to evaluate the models, experimental result reports that enhanced features and dual-attention modules reduce error of fitting ~5% and ~11%, respectively.

A New Prediction Method for Chaotic Time Series

2013 ◽  
Vol 380-384 ◽  
pp. 1673-1676
Author(s):  
Juan Du
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Prediction Accuracy ◽  
Time Series Prediction ◽  
Prediction Method ◽  
Cumulative Effect ◽  
Training Data ◽  
Training Algorithm ◽  
Learning Speed ◽  
Sunspot Data

In order to show the time cumulative effect in the process for the time series prediction, the process neural network is taken. The training algorithm of modified particle swarm is used to the model for the learning speed. The training data is sunspot data from 1700 to 2007. Simulation result shows that the prediction model and algorithm has faster training speed and prediction accuracy than the artificial neural network.

Data-driven stochastic model for cross-interacting processes with different time scales

2021 ◽  
Author(s):  
Andrey Gavrilov ◽  
Aleksei Seleznev ◽  
Dmitry Mukhin ◽  
Alexander Feigin
Keyword(s):  
Time Series ◽  
Stochastic Model ◽  
Time Scales ◽  
Time Scale ◽  
Sampling Interval ◽  
Bayesian Optimization ◽  
Data Driven ◽  
Middle Pleistocene ◽  
Model Based ◽  
Different Time Scales

<p>The problem of modeling interaction between processes with different time scales is very important in geoscience. In this report, we propose a new form of empirical evolution operator model based on the analysis of multiple time series representing processes with different time scales. We assume that the time series are given on the same time interval.</p><p>To construct the model, we extend the previously developed general form of nonlinear stochastic model based on artificial neural networks and designed for the case of time series with constant sampling interval [1]. This sampling interval is related to the main time scale of the process under consideration, which is described by the deterministic component of the model, while the faster time scales are modeled by its stochastic component, possibly depending on the system’s state. This model also includes slower processes in the form of weak time-dependence, as well as external forcing. The structure of the model is optimized using Bayesian approach [1]. The model has proven its efficiency in a number of applications [2-4].</p><p>The idea of modeling time series with different time scales is to formulate the above-described model individually for each time scale, and then to include the parameterized influence of the other time scales in it. Particularly, the influence of “slower” time series is included in the form of parameter trends, and the influence of “faster” time series is included by time-averaging their statistics. The algorithm and first results of comparison between the new model and the model without cross-interactions will be discussed.</p><p>The work was supported by the Russian Science Foundation (Grant No. 20-62-46056).</p><p>1. Gavrilov, A., Loskutov, E., & Mukhin, D. (2017). Bayesian optimization of empirical model with state-dependent stochastic forcing. Chaos, Solitons & Fractals, 104, 327–337. http://doi.org/10.1016/j.chaos.2017.08.032</p><p>2. Mukhin, D., Kondrashov, D., Loskutov, E., Gavrilov, A., Feigin, A., & Ghil, M. (2015). Predicting Critical Transitions in ENSO models. Part II: Spatially Dependent Models. Journal of Climate, 28(5), 1962–1976. http://doi.org/10.1175/JCLI-D-14-00240.1</p><p>3. Gavrilov, A., Seleznev, A., Mukhin, D., Loskutov, E., Feigin, A., & Kurths, J. (2019). Linear dynamical modes as new variables for data-driven ENSO forecast. Climate Dynamics, 52(3–4), 2199–2216. http://doi.org/10.1007/s00382-018-4255-7</p><p>4. Mukhin, D., Gavrilov, A., Loskutov, E., Kurths, J., & Feigin, A. (2019). Bayesian Data Analysis for Revealing Causes of the Middle Pleistocene Transition. Scientific Reports, 9(1), 7328. http://doi.org/10.1038/s41598-019-43867-3</p>

scholarly journals Improving prediction accuracy of river discharge time series using a Wavelet-NAR artificial neural network

2012 ◽  
Vol 14 (4) ◽  
pp. 974-991 ◽  
Author(s):  
Shouke Wei ◽  
Depeng Zuo ◽  
Jinxi Song
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Artificial Neural Network ◽  
Hybrid Model ◽  
River Discharge ◽  
Prediction Accuracy ◽  
Hybrid Approach ◽  
Series Data ◽  
Discharge Time ◽  
Artificial Neural

This study developed a wavelet transformation and nonlinear autoregressive (NAR) artificial neural network (ANN) hybrid modeling approach to improve the prediction accuracy of river discharge time series. Daubechies 5 discrete wavelet was employed to decompose the time series data into subseries with low and high frequency, and these subseries were then used instead of the original data series as the input vectors for the designed NAR network (NARN) with the Bayesian regularization (BR) optimization algorithm. The proposed hybrid approach was applied to make multi-step-ahead predictions of monthly river discharge series in the Weihe River in China. The prediction results of this hybrid model were compared with those of signal NARNs and the traditional Wavelet-Artificial Neural Network hybrid approach (WNN). The comparison results revealed that the proposed hybrid model could significantly increase the prediction accuracy and prediction period of the river discharge time series in the current case study.

scholarly journals A Hybrid Model for Monthly Precipitation Time Series Forecasting Based on Variational Mode Decomposition with Extreme Learning Machine

Information ◽  
2018 ◽  
Vol 9 (7) ◽  
pp. 177 ◽  
Author(s):  
Guohui Li ◽  
Xiao Ma ◽  
Hong Yang
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Prediction Model ◽  
Hybrid Model ◽  
Prediction Accuracy ◽  
Monthly Precipitation ◽  
Variational Mode Decomposition ◽  
Precipitation Time Series ◽  
Mode Decomposition ◽  
Learning Machine

The matter of success in forecasting precipitation is of great significance to flood control and drought relief, and water resources planning and management. For the nonlinear problem in forecasting precipitation time series, a hybrid prediction model based on variational mode decomposition (VMD) coupled with extreme learning machine (ELM) is proposed to reduce the difficulty in modeling monthly precipitation forecasting and improve the prediction accuracy. The monthly precipitation data in the past 60 years from Yan’an City and Huashan Mountain, Shaanxi Province, are used as cases to test this new hybrid model. First, the nonstationary monthly precipitation time series are decomposed into several relatively stable intrinsic mode functions (IMFs) by using VMD. Then, an ELM prediction model is established for each IMF. Next, the predicted values of these components are accumulated to obtain the final prediction results. Finally, three predictive indicators are adopted to measure the prediction accuracy of the proposed hybrid model, back propagation (BP) neural network, Elman neural network (Elman), ELM, and EMD-ELM models: mean absolute error (MAE), root mean squared error (RMSE), and mean absolute percentage error (MAPE). The experimental simulation results show that the proposed hybrid model has higher prediction accuracy and can be used to predict the monthly precipitation time series.

A Novel Dual Path Gated Recurrent Unit Model for Sea Surface Salinity Prediction

2020 ◽  
Vol 37 (2) ◽  
pp. 317-325 ◽  
Author(s):  
Tao Song ◽  
Zihe Wang ◽  
Pengfei Xie ◽  
Nisheng Han ◽  
Jingyu Jiang ◽  
...  
Keyword(s):  
Neural Network ◽  
Time Series ◽  
South China Sea ◽  
South China ◽  
Prediction Accuracy ◽  
Sea Surface ◽  
Sea Surface Salinity ◽  
Surface Salinity ◽  
Intrinsic Correlation ◽  
Gated Recurrent Unit

AbstractAccurate and real-time sea surface salinity (SSS) prediction is an elemental part of marine environmental monitoring. It is believed that the intrinsic correlation and patterns of historical SSS data can improve prediction accuracy, but they have been not fully considered in statistical methods. In recent years, deep-learning methods have been successfully applied for time series prediction and achieved excellent results by mining intrinsic correlation of time series data. In this work, we propose a dual path gated recurrent unit (GRU) network (DPG) to address the SSS prediction accuracy challenge. Specifically, DPG uses a convolutional neural network (CNN) to extract the overall long-term pattern of time series, and then a recurrent neural network (RNN) is used to track the local short-term pattern of time series. The CNN module is composed of a 1D CNN without pooling, and the RNN part is composed of two parallel but different GRU layers. Experiments conducted on the South China Sea SSS dataset from the Reanalysis Dataset of the South China Sea (REDOS) show the feasibility and effectiveness of DPG in predicting SSS values. It achieved accuracies of 99.29%, 98.44%, and 96.85% in predicting the coming 1, 5, and 14 days, respectively. As well, DPG achieves better performance on prediction accuracy and stability than autoregressive integrated moving averages, support vector regression, and artificial neural networks. To the best of our knowledge, this is the first time that data intrinsic correlation has been applied to predict SSS values.

scholarly journals Multi-scale event synchronization analysis for unravelling climate processes: A wavelet-based approach

2017 ◽  
Author(s):  
Ankit Agarwal ◽  
Norbert Marwan ◽  
Maheswaran Rathinasamy ◽  
Bruno Merz ◽  
Jürgen Kurths
Keyword(s):  
Time Series ◽  
Time Scales ◽  
Time Scale ◽  
Temporal Dynamics ◽  
Multiple Scales ◽  
Nonlinear Interactions ◽  
Reference Time ◽  
Multi Scale ◽  
Spatio Temporal ◽  
Different Time Scales

Abstract. The temporal dynamics of climate processes are spread across different time scales and, as such, the study of these processes only at one selected time scale might not reveal the complete mechanisms and interactions within and between the (sub-) processes. For capturing the nonlinear interactions between climatic events, the method of event synchronization has found increasing attention recently. The main drawback with the present estimation of event synchronization is its restriction to analyse the time series at one reference time scale only. The study of event synchronization at multiple scales would be of great interest to comprehend the dynamics of the investigated climate processes. In this paper, wavelet based multi-scale event synchronization (MSES) method is proposed by combining the wavelet transform and event synchronization. Wavelets are used extensively to comprehend multi-scale processes and the dynamics of processes across various time scales. The proposed method allows the study of spatio-temporal patterns across different time scales. The method is tested on synthetic and real-world time series in order to check its replicability and applicability. The results indicate that MSES is able to capture relationships that exist between processes at different time scales.

Sign in / Sign up

Export Citation Format

Share Document