Decomposition aided attention-based recurrent neural networks for multistep ahead time-series forecasting of renewable power generation

Renewable energy plays an increasingly important role in our future. As fossil fuels become more difficult to extract and effectively process, renewables offer a solution to the ever-increasing energy demands of the world. However, the shift toward renewable energy is not without challenges. While fossil fuels offer a more reliable means of energy storage that can be converted into usable energy, renewables are more dependent on external factors used for generation. Efficient storage of renewables is more difficult often relying on batteries that have a limited number of charge cycles. A robust and efficient system for forecasting power generation from renewable sources can help alleviate some of the difficulties associated with the transition toward renewable energy. Therefore, this study proposes an attention-based recurrent neural network approach for forecasting power generated from renewable sources. To help networks make more accurate forecasts, decomposition techniques utilized applied the time series, and a modified metaheuristic is introduced to optimized hyperparameter values of the utilized networks. This approach has been tested on two real-world renewable energy datasets covering both solar and wind farms. The models generated by the introduced metaheuristics were compared with those produced by other state-of-the-art optimizers in terms of standard regression metrics and statistical analysis. Finally, the best-performing model was interpreted using SHapley Additive exPlanations.

A new framework for water quality forecasting coupling causal inference, time-frequency analysis and uncertainty quantification

Accurate forecasting of water quality variables in river systems is crucial for relevant administrators to identify potential water quality degradation issues and take countermeasures promptly. However, pure data-driven forecasting models are often insufficient to deal with the highly varying periodicity of water quality in today's more complex environment. This study presents a new holistic framework for time-series forecasting of water quality parameters by combining advanced deep learning algorithms (i.e., Long Short-Term Memory (LSTM) and Informer) with causal inference, time-frequency analysis, and uncertainty quantification. The framework was demonstrated for total nitrogen (TN) forecasting in the largest artificial lakes in Asia (i.e., the Danjiangkou Reservoir, China) with six-year monitoring data from January 2017 to June 2022. The results showed that the pre-processing techniques based on causal inference and wavelet decomposition can significantly improve the performance of deep learning algorithms. Compared to the individual LSTM and Informer models, wavelet-coupled approaches diminished well the apparent forecasting errors of TN concentrations, with 24.39%, 32.68%, and 41.26% reduction at most in the average, standard deviation, and maximum values of the errors, respectively. In addition, a post-processing algorithm based on the Copula function and Bayesian theory was designed to quantify the uncertainty of predictions. With the help of this algorithm, each deterministic prediction of our model can correspond to a range of possible outputs. The 95% forecast confidence interval covered almost all the observations, which proves a measure of the reliability and robustness of the predictions. This study provides rich scientific references for applying advanced data-driven methods in time-series forecasting tasks and a practical methodological framework for water resources management and similar projects.

Analysis of the effectiveness of non-pharmaceutical interventions on influenza during the Coronavirus disease 2019 pandemic by time-series forecasting

BACKGROUND

Coronavirus disease 2019 (COVID-19) was first identified in South Korea during the 2019-2020 seasonal influenza epidemic. The social distancing measures, as effective non-pharmaceutical interventions (NPIs), adopted to mitigate the spread of COVID-19 might have influenced influenza activity. We evaluated IFV(influenza virus) activity during the COVID-19 pandemic and the effect of NPI intensity on influenza transmission.

METHODS

IFV activity and epidemic duration during COVID-19 pandemic were predicted under a counterfactual scenario with no NPIs against COVID-19. The Seasonal Autoregressive Integrated Moving Average Model was used to quantify the effects of NPIs on the transmission of influenza virus. Influenza-like illness/1000 outpatients and IFV positivity rate from the 2011-2012 to 2021-2022 seasons were used in this study.

RESULTS

Comparison of the 2020-2021 and 2021-2022 seasonal influenza activities with those in 2013-2019 showed that COVID-19 outbreaks and associated NPIs such as face mask use, school closures, and travel restrictions reduced the influenza incidence by 91%. Without NPIs against COVID-19, the rates of influenza-like illness and IFV positivity would have been high during the influenza epidemic season, as in previous seasons. NPI intensity decreased the transmission of influenza; the magnitude of the reduction increased as the intensity of social-distancing measures increased (weak social distancing; step-by-step daily recovery: 58.10%, strong social distancing; special quarantine measures: 95.12%).

CONCLUSIONS

Our results suggest that NPIs and personal hygiene can be used to suppress influenza transmission. NPIs against COVID-19 may be useful strategies for the prevention and control of influenza epidemics.

Model-free forecasting of partially observable spatiotemporally chaotic systems

Reservoir computing is a powerful tool for forecasting turbulence because its simple architecture has the computational efficiency to handle high-dimensional systems. Its implementation, however, often requires full state-vector measurements and knowledge of the system nonlinearities. We use nonlinear projector functions to expand the system measurements to a high dimensional space and then feed them to a reservoir to obtain forecasts. We demonstrate the application of such reservoir computing networks on spatiotemporally chaotic systems, which model several features of turbulence. We show that using radial basis functions as nonlinear projectors enables complex system nonlinearities to be captured robustly even with only partial observations and without knowing the governing equations. Finally, we show that when measurements are sparse or incomplete and noisy, such that even the governing equations become inaccurate, our networks can still produce reasonably accurate forecasts, thus paving the way towards model-free forecasting of practical turbulent systems.

Data-driven prediction of greenhouse aquaponics air temperature based on adaptive time pattern network

Greenhouse aquaponics system (GHAP) improves productivity by harmonizing internal environments. Keeping a suitable air temperature of GHAP is essential for the growth of plant and fish. However, the disturbance of various environmental factors and the complexity of temporal patterns affect the accuracy of the microclimate time-series forecasting. This work proposed an Adaptive Time Pattern Network (ATPNet) to predict GHAP air temperature, which consists of deep temporal feature (DTF) module, multiple temporal pattern convolution (MTPC) module, and spatial attention mechanism (SAM) module. The DTF module has a wide sensory range and can capture information over a long-time span. The MTPC module is designed to improve model response performance by exploiting the effective temporal information of different environmental factors at different times. At the same time, the SAM can explore the correlations among different environmental factors. The ATPNet found that air temperature of GHAP has a strong correlation with other temperature-related parameters (external air temperature, external soil temperature, and water temperature). Compared with the best performance of three baseline models (multilayer perceptron (MLP), recurrent neural network (RNN), and Temporal Convolutional Network (TCN)), the ATPNet enhanced overall prediction performance for the following 24 h by 7.44% for root mean squared error (RMSE), 2.53% for mean absolute error (MAE), and 3.15% for mean absolute percentage error (MAPE), respectively.

Spatial-Temporal Synchronous Graph Transformer network (STSGT) for COVID-19 forecasting

COVID-19 has become a matter of serious concern over the last few years. It has adversely affected numerous people around the globe and has led to the loss of billions of dollars of business capital. In this paper, we propose a novel Spatial-Temporal Synchronous Graph Transformer network (STSGT) to capture the complex spatial and temporal dependency of the COVID-19 time series data and forecast the future status of an evolving pandemic. The layers of STSGT combine the graph convolution network (GCN) with the self-attention mechanism of transformers on a synchronous spatial-temporal graph to capture the dynamically changing pattern of the COVID time series. The spatial-temporal synchronous graph simultaneously captures the spatial and temporal dependencies between the vertices of the graph at a given and subsequent time-steps, which helps capture the heterogeneity in the time series and improve the forecasting accuracy. Our extensive experiments on two publicly available real-world COVID-19 time series datasets demonstrate that STSGT significantly outperforms state-of-the-art algorithms that were designed for spatial-temporal forecasting tasks. Specifically, on average over a 12-day horizon, we observe a potential improvement of 12.19% and 3.42% in Mean Absolute Error (MAE) over the next best algorithm while forecasting the daily infected and death cases respectively for the 50 states of US and Washington, D.C. Additionally, STSGT also outperformed others when forecasting the daily infected cases at the state level, e.g., for all the counties in the State of Michigan. The code and models are publicly available at https://github.com/soumbane/STSGT.

Historical and future trends in emergency pituitary referrals: a machine learning analysis

PURPOSE

Acute pituitary referrals to neurosurgical services frequently necessitate emergency care. Yet, a detailed characterisation of pituitary emergency referral patterns, including how they may change prospectively is lacking. This study aims to evaluate historical and current pituitary referral patterns and utilise state-of-the-art machine learning tools to predict future service use.

METHODS

A data-driven analysis was performed using all available electronic neurosurgical referrals (2014-2021) to the busiest U.K. pituitary centre. Pituitary referrals were characterised and volumes were predicted using an auto-regressive moving average model with a preceding seasonal and trend decomposition using Loess step (STL-ARIMA), compared against a Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) algorithm, Prophet and two standard baseline forecasting models. Median absolute, and median percentage error scoring metrics with cross-validation were employed to evaluate algorithm performance.

RESULTS

462 of 36,224 emergency referrals were included (referring centres = 48; mean patient age = 56.7 years, female:male = 0.49:0.51). Emergency medicine and endocrinology accounted for the majority of referrals (67%). The most common presentations were headache (47%) and visual field deficits (32%). Lesions mainly comprised tumours or haemorrhage (85%) and involved the pituitary gland or fossa (70%). The STL-ARIMA pipeline outperformed CNN-LSTM, Prophet and baseline algorithms across scoring metrics, with standard accuracy being achieved for yearly predictions. Referral volumes significantly increased from the start of data collection with future projected increases (p < 0.001) and did not significantly reduce during the COVID-19 pandemic.

CONCLUSION

This work is the first to employ large-scale data and machine learning to describe and predict acute pituitary referral volumes, estimate future service demands, explore the impact of system stressors (e.g. COVID pandemic), and highlight areas for service improvement.

Impact of the COVID-19 pandemic on road freight transportation - A Colombian case study

The COVID-19 pandemic has wreaked havoc worldwide, with profound economic, environmental, and social implications. Fears about the economic situation have called attention to freight transportation performance as a derivative economic development demand. It is critical to economic growth, especially in industrialized countries with a strong positive relationship to road transport. The empirical evidence in developing countries, such as Colombia, showed that economic growth has had linked to road freight transport regardless of comparisons with other economic sectors. Thie main objective of the paper is to assess the impact caused by the COVID-19 restrictions on freight transport by road to depict the challenge between freight transport performance and economic growth. The total freight transported in 2020 was predicted based on a time series analysis considering the system's performance. As a result, in 2020, the ton of freight transported became only 40% of the predicted freight in the most critical month. The analysis can help planners implement policies to improve freight transport behavior and react during future economic downturns. Events such as the COVID-19 pandemic demonstrate that freight transportation must be fast and flexible, is crucial to act in the short term, and consider the long-term recovery. In addition, cooperation among the various economic sectors' stakeholders must be necessary.

GLYFE: review and benchmark of personalized glucose predictive models in type 1 diabetes

Due to the sensitive nature of diabetes-related data, preventing them from being easily shared between studies, and the wide discrepancies in their data processing pipeline, progress in the field of glucose prediction is hard to assess. To address this issue, we introduce GLYFE (GLYcemia Forecasting Evaluation), a benchmark of machine learning-based glucose predictive models. We present the accuracy and clinical acceptability of nine different models coming from the literature, from standard autoregressive to more complex neural network-based models. These results are obtained on two different datasets, namely UVA/Padova Type 1 Diabetes Metabolic Simulator (T1DMS) and Ohio Type-1 Diabetes Mellitus (OhioT1DM), featuring artificial and real type 1 diabetic patients respectively. By providing extensive details about the data flow as well as by providing the whole source code of the benchmarking process, we ensure the reproducibility of the results and the usability of the benchmark by the community. Those results serve as a basis of comparison for future studies. In a field where data are hard to obtain, and where the comparison of results from different studies is often irrelevant, GLYFE gives the opportunity of gathering researchers around a standardized common environment.

Identifying the impact of the COVID-19 pandemic on driving behavior using naturalistic driving data and time series forecasting

INTRODUCTION

COVID-19 has disrupted daily life and societal flow globally since December 2019; it introduced measures such as lockdown and suspension of all non-essential movements. As a result, driving activity was also significantly affected. Still, to-date, a quantitative assessment of the effect of COVID-19 on driving behavior during the lockdown is yet to be provided. This gap forms the motivation for this paper, which aims at comparing observed values concerning three indicators (average speed, speeding, and harsh braking), with forecasts based on their corresponding observations before the lockdown in Greece.

METHOD

Time series of the three indicators were extracted using a specially developed smartphone application and transmitted to a back-end platform between 01/01/2020 and 09/05/2020, a time period containing normal operations, COVID-19 spreading, and the full lockdown period in Greece. Based on the collected data, XGBoost was employed to identify the most influential COVID-19 indicators, and Seasonal AutoRegressive Integrated Moving Average (SARIMA) models were developed for obtaining forecasts on driving behavior.

RESULTS

Results revealed the intensity of the impact of COVID-19 on driving, especially on average speed, speeding, and harsh braking per 100 km. More specifically, speeds were found to increase by 2.27 km/h on average compared to the forecasted evolution, while harsh braking/100 km increased to almost 1.51 on average. On the bright side, road crashes in Greece were reduced by 49% during the months of COVID-19 compared to the non-COVID-19 period.

Long-term prediction for temporal propagation of seasonal influenza using Transformer-based model

Influenza is one of the most common infectious diseases worldwide, which causes a considerable economic burden on hospitals and other healthcare costs. Predicting new and urgent trends in epidemiological data is an effective way to prevent influenza outbreaks and protect public health. Traditional autoregressive(AR) methods and new deep learning models like Recurrent Neural Network(RNN) have been actively studied to solve the problem. Most existing studies focus on the short-term prediction of influenza. Recently, Transformer models show superior performance in capturing long-range dependency than RNN models. In this paper, we develop a Transformer-based model, which utilizes the potential of the Transformer to increase the prediction capacity. To fuse information from data of different sources and capture the spatial dependency, we design a sources selection module based on measuring curve similarity. Our model is compared with the widely used AR models and RNN-based models on USA and Japan datasets. Results show that our approach provides approximate performance in short-term forecasting and better performance in long-term forecasting.

Performance Evaluation of Soft Computing Approaches for Forecasting COVID-19 Pandemic Cases

An unexpected outbreak of deadly Covid-19 in later part of 2019 not only endangered the economies of the world but also posed threats to the cultural, social and psychological barriers of mankind. As soon as the virus emerged, scientists and researchers from all over the world started investigating the dynamics of this disease. Despite extensive investments in research, no cure has been officially found to date. This uncertain situation rises severe threats to the survival of mankind. An ultimate need of the time is to investigate the course of disease transfer and suggest a future projection of the disease transfer to be enabled to effectively tackle the always evolving situations ahead. In the present study daily new cases of COVID-19 was predicted using different forecasting techniques; Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing/Error Trend Seasonality (ETS), Artificial Neural Network Models (ANN), Gene Expression Programming (GEP), and Long Short-Term Memory (LSTM) in four countries; Pakistan, USA, India and Brazil. The dataset of new daily confirmed cases of COVID-19 from the date on which first case was registered in the respective country to 30 November 2020 is analyzed through these five forecasting models to forecast the new daily cases up to 31st January 2020. The forecasting efficiency of each model was evaluated using well known statistical parameters R 2, RMSE, and NSE. A comparative analysis of all above-mentioned models was performed. Finally, the study concluded that Long Short-Term Memory (LSTM) neural network-based forecasting model projected the future cases of COVID-19 pandemic best in all the selected four stations. The accuracy of the model ranges from coefficient of determination value of 0.85 in Brazil to 0.96 in Pakistan. NSE value for the model in India is 0. 99, 0.98 in USA and Pakistan and 0.97 in Brazil. This high-accuracy forecast of COVID-19 cases enables the projection of possible peaks in near future in the aforementioned countries and, therefore, prove to be helpful in formulating strategies to get prepared for the potential hard times ahead.

Time-series forecasting of Bitcoin prices using high-dimensional features: a machine learning approach

Bitcoin is a decentralized cryptocurrency, which is a type of digital asset that provides the basis for peer-to-peer financial transactions based on blockchain technology. One of the main problems with decentralized cryptocurrencies is price volatility, which indicates the need for studying the underlying price model. Moreover, Bitcoin prices exhibit non-stationary behavior, where the statistical distribution of data changes over time. This paper demonstrates high-performance machine learning-based classification and regression models for predicting Bitcoin price movements and prices in short and medium terms. In previous works, machine learning-based classification has been studied for an only one-day time frame, while this work goes beyond that by using machine learning-based models for one, seven, thirty and ninety days. The developed models are feasible and have high performance, with the classification models scoring up to 65% accuracy for next-day forecast and scoring from 62 to 64% accuracy for seventh-ninetieth-day forecast. For daily price forecast, the error percentage is as low as 1.44%, while it varies from 2.88 to 4.10% for horizons of seven to ninety days. These results indicate that the presented models outperform the existing models in the literature.