Time Series Methods Research Articles

Forecasting short-term subway passenger flow using Wi-Fi data: comparative analysis of advanced time-series methods

Forecasting short-term subway passenger flow using Wi-Fi data: comparative analysis of advanced time-series methods

Development of New Electricity System Marginal Price Forecasting Models Using Statistical and Artificial Intelligence Methods

The System Marginal Price (SMP) is the cost of the last unit of electricity supplied to the grid, reflecting the supply–demand equilibrium and serving as a key indicator of market conditions. Accurate SMP forecasting is essential for ensuring market stability and economic efficiency. This study addresses the challenges of SMP prediction in Turkey by proposing a comprehensive forecasting framework that integrates machine learning, deep learning, and statistical models. Advanced feature selection techniques, such as Minimum Redundancy Maximum Relevance (mRMR) and Maximum Likelihood Feature Selector (MLFS), are employed to refine model inputs. The framework incorporates time series methods like Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), and Convolutional LSTM (ConvLSTM) to capture complex temporal patterns, alongside models such as Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Extreme Learning Machine (ELM) for modeling non-linear relationships. Model performance was evaluated using the Mean Absolute Percentage Error (MAPE) across regular weekdays, weekends, and public holidays. XGBoost combined with MLFS consistently achieved the lowest MAPE values, demonstrating exceptional accuracy and robustness. Among all of the models, XGBoost combined with MLFS consistently achieved the lowest MAPE values, demonstrating superior accuracy and robustness. The results highlight the inadequacy of traditional models like ARIMA and SARIMA in capturing non-linear and highly volatile patterns, reinforcing the necessity of using advanced techniques for effective SMP forecasting. Overall, this study presents a novel and comprehensive approach tailored for complex electricity markets, significantly enhancing predictive reliability by incorporating economic indicators and sophisticated feature selection methods.

Physics-informed ensemble learning with residual modeling for enhanced building energy prediction

Accurate modeling of building energy use is important to support a diverse spectrum of its downstream applications, such as building energy efficiency assessment, resilience analysis, smart control, etc. As mainstream approaches of building energy modeling, physics-based modeling builds on different fidelities of physics rules yet are usually compromised in modeling accuracy due to insufficiency of physics rules to capture real-world dynamics and incomplete input information. Data-driven approaches are computationally efficient, but black box (uninterpretable) in nature. For improved modeling of building energy use, this work proposes a physics-informed ensemble learning approach in building energy prediction through residual modeling. Specifically, we first analyze the components of building energy use data. Evidence suggests that the building energy use data can be decomposed into the physics-driven part, occupant-driven part, and white noise. Second, high-fidelity physics-based building models (EnergyPlus) and low-fidelity ones (RC models) are developed to capture the physics-driven part while time series methods are explored as the residual modeling approach to capture the occupant-driven discrepancies between physics-based simulation and measured building energy use (i.e., residuals). Finally, the physics-informed ensemble learning is proposed to integrate physics-based and data-driven models for enhanced accuracy of building energy modeling. Results demonstrate 40–90% decrease of discrepancies between modeling and field observations compared to traditional physics-based modeling methods. Moreover, when the training dataset size is small, the proposed ensemble model overperforms the pure data-driven models, demonstrating its higher robustness in extrapolation scenarios. This work makes fundamental contributions to the development of convergent modeling approaches in the building modeling field.

Early Fault Detection and Operator-Based MIMO Fault-Tolerant Temperature Control of Microreactor

A microreactor is a chemical reaction device that mixes liquids in a very narrow channel and continuously generates reactions. They are attracting attention as next-generation chemical reaction devices because of their ability to achieve small-scale and highly efficient reactions compared to the conventional badge method. However, the challenge is to design a control system that is tolerant of faults in some of the enormous number of sensors in order to achieve parallel production by numbering up. In a previous study, a simultaneous control system for two different temperatures was proposed in an experimental system that imitated the microreactor cooled by Peltier devices. In addition, a fault-tolerant control system for one area has also been proposed. However, the fault-tolerant control system could not be applied to the control system of two temperatures in the previous study. In this paper, we extend it to a two-input, two-output fault-tolerant control system. We also use a fault detection system that combines ChangeFinder, a time-series data analysis method, and One-Class SVM, an unsupervised learning method. Finally, the effectiveness of the proposed method is confirmed by experiments.

Time Series Forecasting of Thermal Systems Dispatch in Legal Amazon Using Machine Learning

This paper analyzes time series forecasting methods applied to thermal systems in Brazil, specifically focusing on diesel consumption as a key determinant. Recognizing the critical role of thermal systems in ensuring energy stability, especially during low rain seasons, this study employs bagged, boosted, and stacked ensemble learning methods for time series forecasting focusing on exploring consumption patterns and trends. By leveraging historical data, the research aims to predict future diesel consumption within Brazil’s thermal energy sector. Based on the bagged ensemble learning approach a mean absolute percentage error of 0.089% and a coefficient of determination of 0.9752 were achieved (average considering 50 experiments), showing it to be a promising model for the short-time forecasting of thermal dispatch for the electric power generation system. The bagged model results were better than for boosted and stacked ensemble learning methods, long short-term memory networks, and adaptive neuro-fuzzy inference systems. Since the thermal dispatch in Brazil is closely related to energy prices, the predictions presented here are an interesting way of planning and decision-making for energy power systems.

Models of the Impact of Socio-Economic Shocks on Higher Education Development

This article is devoted to the analysis of the impact of socio-economic shocks on the dynamics of higher education development. It is substantiated that, on the one hand, higher education influences the development of society and the economy, on the other hand, the development trends of a country provide both opportunities and limitations for its development. An algorithmic model for studying the impact of social and economic shocks on the development of the higher education system (HES) has been developed. To diagnose the relationship between higher education and the socio-economic development of Ukraine and Slovakia, the following indicators were used: GDP per capita, the Human Development Index, school enrollment, tertiary, and net migration. The presence of nonlinear trends in the change in indicators has been shown and portraits of the socio-economic development of the countries have been constructed. To assess the impact of socio-economic shocks on the HES, the time-series decomposition method and cross-spectral analysis were used. The time-series decomposition allowed us to identify cyclical components of indicators, based on applying cross-spectral analysis, and the most significant local harmonics and the lag of their influence on the occurrence of shocks in the HES were determined. The use of the developed models allows us to predict periods of shock points in the HES depending on shocks in the tendencies of GDP per capita and net migration.

The Secondary Sex Ratio and Male Mortality at Pre-Reproductive Ages: A Test of Selection In Utero.

The secondary sex ratio (i.e., the ratio of male to female live births; hereafter referred to as the SSR) falls in populations encountering ambient stressors. Much theory and some empirical work indicates that males born to low SSR cohorts may be "positively selected" in that excess culling in utero may correspond with greater than expected survival among live-born males in that cohort. We extend prior work by testing, in historical Utah, whether the SSR varies positively with male mortality at pre-reproductive ages. This study uses detailed records from the Utah Population Database to focus on Utahns born 1850-1940. We use rigorous time-series methods, which control for strong secular declines in mortality as well as ambient perturbations shared equally among males and females, to investigate the male culling inference. We observe a positive relation between the SSR and male mortality during youth (i.e., 5 to < 20 years; p < 0.05) but not in infancy or early childhood. In this historical population, the SSR appears to gauge hardiness of surviving male cohorts. However, whether the high fertility and/or family structure context of Latter-day Saints in historical Utah explains the age-specific pattern of male mortality warrants further scrutiny.

Kidney replacement therapy trends in end-stage kidney disease patients in South Korea during the COVID-19 pandemic.

The corona virus disease 2019 posed a major risk for end-stage kidney disease (ESKD) cases. Our study aimed to assess changes in kidney replacement therapy (KRT) trends and healthcare access for these patients during the pandemic. We retrospectively analyzed nationwide data from July 2017 to June 2022 to assess changes in KRT and ESKD incidence. KRT modalities included peritoneal dialysis (PD), hemodialysis (HD), and kidney transplantation (KT). We utilized the interrupted time series (ITS) method to compare changes in KRT modality before and after the incidence of the COVID-19 pandemic. ESKD incidence remained stable from 2018 to April 2022. The ITS analysis confirmed that the pandemic did not significant impact overall KRT incidence. PD cases decreased (5.7% to 1.3%), while HD cases increased (81.6% to 85%), and KT recipient remained relatively stable (12.7% to 17.3%). The hospitalization and hospital stay decreased in nursing hospital (165.01 days to 147.77 days) and general hospital (61.34 days to 55.58 days) during the pandemic, however, remained unchanged for PD and KT. Our findings indicate no significant changes in ESKD incidence in South Korea during the pandemic. However, there were shifts in modality distribution, with decreased PD and increased HD cases. Notably, HD cases showed a significant reduction in hospital admissions and length of stay. The healthcare system demonstrated stability during the pandemic, with minimal disruptions in ESKD care.

Addressing extreme weather events for the renewable power-water-heating sectors in Neom, Saudi Arabia

A renewable energy design optimization model is proposed to plan investments in power, water, and heat technologies. The intermittent nature of renewables requires that these models capture the variability and complementarity of resources at high spatial and temporal resolutions. However, most planning models use time-series reduction methods that, while capturing data variance, often smooth out extreme weather or demand patterns. To account for extreme patterns and design reliable energy systems, we propose a clustering-optimization framework that considers extreme weather days. This framework is applied to design an integrated multi-sector energy system for the Neom region in Saudi Arabia. Our results show that fully renewable systems designed without considering extreme days could not meet demands and instead required external power or water supplies during a post-optimization simulation. Once extreme days were considered in the optimization, system reliability increased at the expense of larger generation and storage capacity investments.

Tornado Occurrence in the United States as Modulated by Multidecadal Oceanic Oscillations Using Empirical Model Decomposition

Studies have analyzed U.S. tornado variability and correlated F1–F5 tornado occurrence with various natural climate oscillations and anthropogenic factors. Using a relatively new empirical mode decommission (EMD) method that extracts time-frequency modes adaptively without priori assumptions like traditional time-series analysis methods, this study decomposes U.S. tornado variability during 1954–2022 into intrinsic modes on specific temporal scales. Correlating the intrinsic mode functions (IMFs) of EMD with climate indices found that 1. the U.S. overall tornado count is negatively (positively) correlated with the Atlantic Multidecadal Oscillation (AMO) index (the Southern Oscillation Index (SOI)); 2. the negative (positive) correlation tends to be more prevalent in the western (eastern) U.S.; 3. the increase in weak (F1–F2) and decrease in strong (F3–F5) tornadoes after around 2000, when both the AMO and the Pacific Decadal Oscillation (PDO) shifted phases, are likely related to their secular trends and low-frequency IMFs; and 4. the emerging Dixie Tornado Alley coincides with an amplifying intrinsic mode of the SOI that correlates positively with the eastern U.S. and Dixie Alley tornadoes. The long-term persistence of these climate indices can offer potential guidance for future planning for tornado hazards.

A study of short-term wind power segmentation forecasting method considering weather on ramp segments

The short-term fluctuation of wind power can affect its prediction accuracy. Thus, a short-term segmentation prediction method of wind power based on ramp segment division is proposed. A time-series trend extraction method based on moving average iteration is proposed on the full-time period to analyze the real-time change characteristics of power time-series initially; secondly, a ramp segment extraction method based on its definition and identification technique is proposed based on the results of the trend extraction; and a segmentation prediction scheme is proposed to lean the power prediction under different time-series: the LightGBM-LSTM is proposed for the non-ramping segment using point prediction, and the CNN-BiGRU-KDE is proposed for probabilistic prediction of ramp segments. From the results, this ramp segment definition and identification technique can effectively identify the ramp process of wind power, which makes up for the misidentification and omission of the classical climbing event definition; meanwhile, the segment prediction scheme not only meets the prediction accuracy requirements of the non-ramping segment, but also provides the effective robust information for the prediction of the ramping period, which offers reliable reference information for the actual wind farms. In particular, it is well adapted to wind power prediction under extreme working conditions caused by ramping weather, which is a useful addition to short-term wind power prediction research.

Spatio-Temporal Model to Forecast COVID-19 Confirmed Cases in High-Density Areas of Malaysia

The coronavirus 2019 disease has spread across the world. The number of coronaviruses 2019 (COVID-19) cases throughout Malaysia is high in the densely populated state of Selangor. In assisting the early preventive measures, this study utilises time series methods to model and forecast the number of daily positive cases in three Selangor districts: Petaling, Hulu Langat, and Klang. Specifically, the study compares the effectiveness of the Autoregressive Integrated Moving Average (ARIMA), a univariate model and the Generalized Space-Time autoregressive integrated (GSTARI), a multivariate model. For the GSTARI model, uniform and inverse distance weights represent the relationship between locations. The analysed data are from January to August 2021, and the lowest root mean square error (RMSE) is chosen as the best model. The results show GSTARI (1,1) with both spatial weights outperformed ARIMA (0,1,1) in Petaling and Klang but not in Hulu Langat. However, the average RMSE values show that the most accurate and effective for forecasting the number of daily confirmed positive cases in Selangor is using GSTARI. In conclusion, by utilising advanced time series methods such as spatial analysis, this study provides important insights into forecasting trends of infectious diseases like COVID-19 and can help in early preventive measures.

Causal Learning: Monitoring Business Processes Based on Causal Structures.

Conventional methods for process monitoring often fail to capture the causal relationships that drive outcomes, making hard to distinguish causal anomalies from mere correlations in activity flows. Hence, there is a need for approaches that allow causal interpretation of atypical scenarios (anomalies), allowing to identify the influence of operational variables on these anomalies. This article introduces (CaProM), an innovative technique based on causality techniques, applied during the planning phase in business process environments. The technique combines two causal perspectives: anomaly attribution and distribution change attribution. It has three stages: (1) process events are collected and recorded, identifying flow instances; (2) causal learning of process activities, building a directed acyclic graphs (DAGs) represent dependencies among variables; and (3) use of DAGs to monitor the process, detecting anomalies and critical nodes. The technique was validated with a industry dataset from the banking sector, comprising 562 activity flow plans. The study monitored causal structures during the planning and execution stages, and allowed to identify the main factor behind a major deviation from planned values. This work contributes to business process monitoring by introducing a causal approach that enhances both the interpretability and explainability of anomalies. The technique allows to understand which specific variables have caused an atypical scenario, providing a clear view of the causal relationships within processes and ensuring greater accuracy in decision-making. This causal analysis employs cross-sectional data, avoiding the need to average multiple time instances and reducing potential biases, and unlike time series methods, it preserves the relationships among variables.

The Application of Fuzzy Time Series Method for Web-Based Prediction of Household Chemical Product Sales Stock

The seasonal and trend factors of the market demand greatly influence the sales of household chemical products. Therefore, PT. XYZ company requires an accurate sales prediction system to optimize production and inventory. This research aims to develop a web-based sales prediction system for household chemical products using fuzzy time series method. This method considers the factors that influence sales and generates a prediction model that can assist the company in decision-making. This method allows the use of uncertain and unstructured variables in the prediction. The research collected historical sales data of household chemical products from January to December 2023 through interviews, observations, and literature studies. The system was designed using a structured approach and PHP programming language. The Fuzzy Time Series method was used for decision-making. The accuracy of the Fuzzy Time Series method for sales data from January to December 2020 was calculated through program simulation and manual methods, resulting in an accuracy rate of 13.71%. The application provides information on the sales of each product, with the hope that users can reduce the accumulation of stock every month.

Literature Review on Stochastic Modeling of Wet and Dry Spells

With increasing concerns about climate change and its profound implications for water resources, stochastic modeling of wet and dry spells has emerged as a critical domain within hydrology and climatology, providing insight into the temporal patterns of precipitation and drought occurrences. This review of the literature synthesizes recent advances in the modeling of wet and dry spells using stochastic methods, focusing on publications from 2000 to 2024. The review examines various modeling approaches, including Markov chain models, mixed probability models, non-homogeneous models, and time series approaches. Key findings indicate that Markov chain models are effective in simulating the sequential occurrence of wet and dry spells, with higher-order variants addressing issues of overdispersion. Mixed probability models excel at representing heterogeneity and extremes in precipitation data, especially in regions with distinct wet and dry seasons. Non-homogeneous models are valuable for understanding the temporal dynamics and irregularities of dry spells, revealing significant shifts in their frequency and duration over time. These can be complemented by time-series methods to highlight long-term trends and seasonal variations. The review underscores the importance of regional specificity in modeling approaches to accurately capture local climatic and geographical variability. Recommendations for future research include the integration of hybrid models that combine the strengths of various approaches to improve predictive precision, focusing on the impacts of climate change, and conducting cross-regional comparisons to generalize findings and enhance the robustness of the models.

The impact of recreational cannabis legalization on cannabis-related acute care events among adults with schizophrenia.

Patients with schizophrenia have a higher risk of cannabis use disorder and may be uniquely affected by the legalization of recreational cannabis. This study examined whether cannabis legalization led to changes in acute care utilization among patients with schizophrenia. Using linked health administrative data, we included adult patients with schizophrenia in Ontario from October 2015 to May 2021 (n = 121,061). We examined the differences in cannabis, psychosis, and mental health-related emergency department (ED) visits over three periods: pre-legalization, legalization of flowers and herbs (phase 1), and legalization of edibles, extracts, and topicals (phase 2) using interrupted time-series methods. Our study found that phase 1 was associated with decreases in cannabis-related, mental health-related, and cannabis + psychosis-related ED visits among the patients with schizophrenia. Notably, an immediate 25.8% (95% CI 13.8-37.6%) decrease in cannabis-related ED visits was observed in men, and an immediate 18.5% decrease in mental health-related ED visits (95% CI 6.0-31.2%) in women. These decreases were also shown in the comparative ITS models, demonstrating that the changes observed were distinct from trends in the general population. However, phase 2 was not associated with any significant changes. Despite higher baseline rates of acute care utilization among patients with schizophrenia, cannabis legalization was associated with significant reductions, particularly during phase 1. Our findings suggest that regulatory measures accompanying legalization could enhance the quality and safety of cannabis products, potentially leading to fewer adverse health outcomes in vulnerable patient populations. Further research is needed to optimize healthcare responses for this vulnerable population.

Improving electric vehicle charging forecasting: A hybrid deep learning approach for probabilistic predictions

AbstractElectric vehicles (EVs) have gained significant attention recently. Despite their advantages, challenges in the power grid, such as providing necessary information for optimal operation, persist. High‐precision forecasting techniques are essential to address the nonlinear and complex behavior of EV charging. A hybrid structure based on deep learning, called LSTLNet, has been proposed. LSTLNet combines convolutional neural networks (CNN), gated recurrent neural networks (GRU), attention mechanisms (AM), and automatic regression (AR) models. This combination improves the deterministic forecasting model and addresses the weaknesses of CNN and GRU. Deterministic prediction, which determines only one point of consumption charge, is prone to error. Therefore, probabilistic forecasting, represented as a probability distribution function (PDF) containing comprehensive statistical information, is preferred. A smooth band limit maximum likelihood (SBLM) estimator is used to indirectly predict the PDF from the data. Comparative results with conventional shallow and deep methods for similar time series forecasting demonstrate the superiority of the proposed method for both deterministic and probabilistic forecasting.

Forecasting Population in an Uncertain World: Approaches, New Uses, and Troubling Limitations

AbstractThe long human lifespan enables long run forecasts of population size and age distribution. New methods include biodemographic research on upper limits to life expectancy and incorporation of early experiences affecting later life mortality such as smoking, obesity, and childhood health shocks. Some fertility forecasts incorporate education and quantum‐tempo insights. Statistical time series and Bayesian methods generate probabilistic forecasts. Yet recent decades have brought surprising changes in the economy, natural environment, and vital rates. In these changing circumstances we need new methods and the increasing use of probabilistic models and Bayesian methods incorporating outside information. The increasing use of microsimulation combined with aggregate forecasting methods is a very promising development enabling more detailed and heterogeneous forecasts. Some new uses of stochastic forecasts are interesting in themselves. Probabilistic mortality forecasts are used in finance and insurance, and a new Longevity Swap industry has been built on them. Random sample paths used to generate stochastic population forecasts can stress‐test public pension designs for fiscal stability and intergenerational equity. Population forecasting a few decades ago was a dull backwater of demographic research, but now it is increasingly important and is full of intellectual and technical challenges.

Time Series Methods to Assess the Impact of Regulatory Action: A Study of UK Primary Care and Hospital Data on the Use of Fluoroquinolones.

To illustrate the interest in using interrupted time series (ITS) methods, this study evaluated the impact of the UK MHRA's March 2019 Risk Minimisation Measures (RMM) on fluoroquinolone usage. Monthly and quarterly fluoroquinolone use incidence rates from 2012 to 2022 were analysed across hospital care (Barts Health NHS Trust), primary care (Clinical Practice Research Datalink (CPRD) Aurum and CPRD GOLD), and linked records from both settings (East Scotland). Rates were stratified by age (19-59 and ≥ 60 years old). Seasonality-adjusted segmented regression and ARIMA models were employed to model quarterly and monthly rates, respectively. Post-RMM, with segmented regression, both age groups in Barts Health experienced nearly complete reductions (> 99%); CPRD Aurum saw 20.19% (19-59) and 19.29% ( 60) reductions; no significant changes in CPRD GOLD; East Scotland had 45.43% (19-59) and 41.47% ( 60) decreases. Slope analysis indicated increases for East Scotland (19-59) and both CPRD Aurum groups, but a decrease for CPRD GOLD's 60; ARIMA detected significant step changes in CPRD GOLD not identified by segmented regression and noted a significant slope increase in Barts Health's 19-59 group. Both models showed no post-modelling autocorrelations across databases, yet Barts Health's residuals were non-normally distributed with non-constant variance. Both segmented regression and ARIMA confirmed the reduction of fluoroquinolones use after RMM across four different UK primary care and hospital databases. Model diagnostics showed good performance in eliminating residual autocorrelation for both methods. However, diagnostics for hospital databases with low incident use revealed the presence of heteroscedasticity and non-normal white noise using both methods.

Explainable Artificial Intelligence methods for financial time series

We consider the problem of developing explainable Artificial Intelligence methods to interpret the results of Artificial Intelligence models for time series data, taking time dependency into account. To this end, we extend the Shapley-Lorenz method, normalised by construction, to rtificial Intelligence for time series, such as neural networks and recurrent neural networks. We illustrate the application of our proposal to a time series of Bitcoin prices, which acts as the response variable, along with time series of classical financial prices, which act as explanatory variables.Three main findings emerge from the analysis. First, recurrent neural networks lead to a better performance, in terms of accuracy and robustness, with respect to classic neural networks. Second, the best performing models indicate that Bitcoin prices are affected mostly by their lagged values, and that their explainability, in terms of classical financial assets, is limited. Third, although limited, the contribution of classical assets to Bitcoin price prediction is well captured by recurrent neural networks.