Time Series Models Research Articles

Use of ARIMA Model for Forecasting Consequences Due to Traffic Crashes in the Kingdom of Saudi Arabia

Times series models are important statistical methods for analysing data recorded at points of time which considers the order of observations. In this study, the Autoregressive Integrated Moving Average (ARIMA) model was used to analyse the consequences of traffic crashes in the Kingdom of Saudi Arabia (KSA) from 2002-2022. Over the study period, there was a decreasing trend in the forecasted number of all types of injuries per 1,000 traffic crashes. Moreover, to check the validity of the fitted model, the actual observations are plotted with predicted values from 2016 to 2022 and showed a nearly equal and exact pattern between the total number of predicted values and the actual data. It is concluded that the ARIMA model is a good fit to forecast the parameter of consequences per 1,000 crashes. The decrease in consequences may be due to preventive or mitigation measures by various organisations in KSA.

Machine learning and deep learning prediction models for time-series: a comparative analytical study for the use case of the UK short-term electricity price prediction

Machine learning and deep learning prediction models for time-series: a comparative analytical study for the use case of the UK short-term electricity price prediction

Predicting rainfall using machine learning, deep learning, and time series models across an altitudinal gradient in the North-Western Himalayas

AbstractPredicting rainfall is a challenging and critical task due to its significant impact on society. Timely and accurate predictions are essential for minimizing human and financial losses. The dependence of approximately 60% of agricultural land in India on monsoon rainfall implies the crucial nature of accurate rainfall prediction. Precise rainfall forecasts can facilitate early preparedness for disasters associated with heavy rains, enabling the public and government to take necessary precautions. In the North-Western Himalayas, where meteorological data are limited, the need for improved accuracy in traditional modeling methods for rainfall forecasting is pressing. To address this, our study proposes the application of advanced machine learning (ML) algorithms, including random forest (RF), support vector regression (SVR), artificial neural network (ANN), and k-nearest neighbour (KNN) along with various deep learning (DL) algorithms such as long short-term memory (LSTM), bi-directional LSTM, deep LSTM, gated recurrent unit (GRU), and simple recurrent neural network (RNN). These advanced techniques hold the potential to significantly improve the accuracy of rainfall prediction, offering hope for more reliable forecasts. Additionally, time series techniques, including autoregressive integrated moving average (ARIMA) and trigonometric, Box-Cox transform, arma errors, trend, and seasonal components (TBATS), are proposed for predicting rainfall across the altitudinal gradients of India’s North-Western Himalayas. This approach can potentially revolutionise how we approach rainfall forecasting, ushering in a new era of accuracy and reliability. The effectiveness and accuracy of the proposed algorithms were assessed using meteorological data obtained from six weather stations at different elevations spanning from 1980 to 2021. The results indicate that DL methods exhibit the highest accuracy in predicting rainfall, as measured by the root mean squared error (RMSE) and mean absolute error (MAE), followed by ML algorithms and time series techniques. Among the DL algorithms, the accuracy order was bi-directional LSTM, LSTM, RNN, deep LSTM, and GRU. For the ML algorithms, the accuracy order was ANN, KNN, SVR, and RF. These findings suggest that altitude significantly affects the accuracy of the models, highlighting the need for additional weather stations in this mountainous region to enhance the precision of rainfall prediction.

A New Frontier in Wind Shear Intensity Forecasting: Stacked Temporal Convolutional Networks and Tree-Based Models Framework

Wind shear presents a considerable hazard to aviation safety, especially during the critical phases of takeoff and landing. Accurate forecasting of wind shear events is essential to mitigate these risks and improve both flight safety and operational efficiency. This paper introduces a hybrid Temporal Convolutional Networks and Tree-Based Models (TCNs-TBMs) framework specifically designed for time series modeling and the prediction of wind shear intensity. The framework utilizes the ability of TCNs to capture intricate temporal patterns and integrates it with the predictive strengths of TBMs, such as Extreme Gradient Boosting (XGBoost), Random Forest (RF), and Categorical Boosting (CatBoost), resulting in robust forecast. To ensure optimal performance, hyperparameter tuning was performed using the Covariance Matrix Adaptation Evolution Strategy (CMA-ES), enhancing predictive accuracy. The effectiveness of the framework is validated through comparative analyses with standalone machine learning models such as XGBoost, RF, and CatBoost. The proposed TCN-XGBoost model outperformed these alternatives, achieving a lower Root Mean Squared Error (RMSE: 1.95 for training, 1.97 for testing), Mean Absolute Error (MAE: 1.41 for training, 1.39 for testing), and Mean Absolute Percentage Error (MAPE: 7.90% for training, 7.89% for testing). Furthermore, the uncertainty analysis demonstrated the model’s reliability, with a lower mean uncertainty (7.14 × 10−8) and standard deviation of uncertainty (6.48 × 10−8) compared to other models. These results highlight the potential of the TCNs-TBMs framework to significantly enhance the accuracy of wind shear intensity predictions, emphasizing the value of advanced time series modeling techniques for risk management and decision-making in the aviation industry. This study highlights the framework’s broader applicability to other meteorological forecasting tasks, contributing to aviation safety worldwide.

Research on Forecasting Sales of Pure Electric Vehicles in China Based on the Seasonal Autoregressive Integrated Moving Average–Gray Relational Analysis–Support Vector Regression Model

Aiming to address the complexity and challenges of predicting pure electric vehicle (EV) sales, this paper integrates a time series model, support vector machine and combined model to forecast EV sales in China. Firstly, a seasonal autoregressive integrated moving average (SARIMA) model was constructed using historical EV sales data, and the model was trained on sales statistics to obtain forecasting results. Secondly, variables that were highly correlated with sales were analyzed using gray relational analysis (GRA) and utilized as input parameters for the support vector regression (SVR) model, which was constructed to optimize sales predictions for EVs. Finally, a combined model incorporating different algorithms was verified against market sales data to explore the optimal sales prediction approach. The results indicate that the SARIMA-GRA-SVR model with the squared prediction error and inverse method achieved the best predictive performance, with MAPE, MAE and RMSE values of 12%, 1.45 and 2.08, respectively. This empirical study validates the effectiveness and superiority of the SARIMA-GRA-SVR model in forecasting EV sales.

Abstract 4136911: Daily Dyadic Symptom Appraisal Dynamics During Care Transitions in Persons with Heart Failure and Their Care Partners

Background: Symptom monitoring after heart failure (HF) hospitalization is a central aspect of patient education. However, appraising symptoms can be difficult for persons with HF (PWHFs), and studies suggest family care partners (CPs) play a key supportive role. The purpose of this study is to examine dyadic HF symptom appraisal dynamics in PWHF-CP dyads during hospital-to-home transitions. Methods: We conducted a daily diary study of dyadic symptom appraisal in which PWHFs and CPs were enrolled during a HF hospitalization. Over 5 weeks, both completed daily assessments of the PWHF’s symptoms (HF Somatic Perception Scale, PROMIS Fatigue 4a). Each dyad’s appraisal dynamics were quantified using a set of N of 1 dyadic autoregressive time series models examining total symptoms, hallmark symptoms (dyspnea, edema) and nonspecific symptoms (early/subtle cluster, fatigue). Models were examined for within-dyad dynamics (within-dyad associations on the same day or on a 1-day lag). Patterns in significant dynamics across dyads were narratively described. Results: This analysis includes n=14 cohabitating dyads, all partnered. On average, PWHFs and CPs were 69.6±5.4 and 67.4±8.0 years of age. The sample was relatively gender balanced (43% of PWHFs and 64% of CPs identified as female), and primarily identified as non-Hispanic (86% of PWHFs, 93% of CPs) and White (93% of PWHFs, 86% of CPs). Most PWHFs had non-ischemic cardiomyopathy (64%) and were NYHA Class III/IV (86%), with 57% HFrEF (mean EF 37%±17%). Symptom appraisals were highly variable within and across dyads (Fig 1). Most (n=11, 79%) had significant dyadic dynamics, but dynamics across dyads varied by symptom type and time structure (Fig 2). Conclusions: The types of symptoms that HF care dyads are able to appraise together and the time structure of appraisals are variable during care transitions. Precision approaches tailored to individual dyads are needed to optimize inclusion of CPs in symptom monitoring interventions.

Influence of Blood Sampling Service Process Reengineering on Medical Services Supply: Quasi-Experimental Study.

Tertiary hospitals in China are confronted with significant challenges due to limited spatial capacity and workforce constraints, leading to saturated allocation of medical resources and restricted growth in medical service provision. The incorporation of digital health into medical service process reengineering (MSPR) marks a pivotal transformation and restructuring of conventional health service delivery models. Specifically, the application of MSPR to blood sampling services processes reengineering (BSSPR) holds promise for substantially enhancing the efficiency and quality of medical services through streamlining and optimizing these procedures. However, the comprehensive impact of BSSPR has been infrequently quantified in existing research. This study aims to investigate the influence of BSSPR on the efficiency and quality of medical services and to elucidate the key informative technological support points underpinning BSSPR. Data were collected from both the new and old laboratory information systems from August 1, 2019, to December 31, 2021. A combination of statistical description, chi-square test, and t test was used to compare check-in time and waiting time of outpatients before and after the implementation of BSSPR. An interrupted time-series design was used to analyze the impact of BSSPR on medical service efficiency and quality, enabling the control of confounding variables, including changes in medical human resources and both long- and short-term temporal trends. BSSPR had an impact on the efficiency and quality of medical services. Notably, there was a significant increase in the number of patients receiving blood sampling services, with a daily service volume increase of ~150 individuals (P=.04). The average waiting time for patients decreased substantially from 29 (SD 36) to 11 (SD 11) minutes, indicating a marked improvement in patient experience. During the peak period, the number of patients receiving blood sampling services per working hour statistically increased from 9.56 to 16.77 (P<.001). The interrupted time-series model results demonstrated a reduction in patients' waiting time by an average of 26.1 (SD 3.8; 95% CI -33.64 to -18.57) minutes. Although there was an initial decline in the number of outpatients admitted following BSSPR implementation, an upward trend was observed over time (β=1.13, 95% CI 0.91-1.36). BSSPR implementation for outpatients not only reduced waiting time and improved patients' experience but also augmented the hospital's capacity to provide medical services. This study's findings offer valuable insights into the potential advantages of BSSPR and underscore the significance of harnessing digital technologies to optimize medical service processes. This research serves as a foundational basis and provides scientific support for the promotion and application of BSSPR in other health care contexts. By continuing to explore and refine the integration of digital technologies in health care, we can further enhance patient outcomes and elevate the overall quality of medical services.

A comparative analysis of variants of machine learning and time series models in predicting women’s participation in the labor force

Labor force participation of Egyptian women has been a chronic economic problem in Egypt. Despite the improvement in the human capital front, whether on the education or health indicators, female labor force participation remains persistently low. This study proposes a hybrid machine-learning model that integrates principal component analysis (PCA) for feature extraction with various machine learning and time-series models to predict women’s employment in times of crisis. Various machine learning (ML) algorithms, such as support vector machine (SVM), neural network, K-nearest neighbor (KNN), linear regression, random forest, and AdaBoost, in addition to popular time series algorithms, including autoregressive integrated moving average (ARIMA) and vector autoregressive (VAR) models, have been applied to an actual dataset from the public sector. The manpower dataset considered gender from different regions, ages, and educational levels. The dataset was then trained, tested, and evaluated. For performance validation, forecasting accuracy metrics were constructed using mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), mean absolute percent error (MAPE), R-squared (R2), and cross-validated root mean squared error (CVRMSE). Another Dickey-Fuller test was performed to evaluate and compare the accuracy of the applied models, and the results showed that AdaBoost outperforms the other methods by an accuracy of 100%. Compared to alternative works, our findings demonstrate a comprehensive comparative analysis for predicting women’s participation in different regions during an economic crisis.

Unveiling an asymmetric relationship between global crude oil and local food prices in an oil-importing economy

AbstractRecent swift comovements of local food and global crude oil prices have attracted the attention of policymakers and researchers. To evaluate this relationship, many studies have used time series models to explore global crude oil and local food prices. However, robust research based on advanced nonlinear time series models that incorporate control variables for their formation is lacking. In this paper, nonlinear techniques are applied to assess the asymmetric nexus between Brent oil prices and local retail food prices in Slovakia. To estimate this value, we extend the single-threshold NARDL approach to the MTNARDL model. The nominal exchange rate and industrial production index are used as the control variables. Compared with conventional NARDL models, the MTNARDL model provides a more detailed representation of global oil‒local food price linkages and detects the asymmetric effect of global oil prices on food prices from both long- and short-term perspectives. Interestingly, with respect to long- and short-term food price volatility, changes in response to oil price fluctuations are greatest under a regime with rather a small number of positive and moderate changes.

PHM-Based Modeling for Cyberattack Classifier Performance

This research implements Prognostics and Health Management (PHM) using multiple linear regression and multivariate time series models to monitor and predict when the performance of a Machine Learning-based cyberattack classifier might degrade to an unacceptable level, enabling preemptive maintenance strategies.

Temporal cross‐validation in forecasting: A case study of COVID‐19 incidence using wastewater data

AbstractTwo predominant methodologies in forecasting temporal processes include traditional time series models and machine learning methods. This paper investigates the impact of time series cross‐validation (TSCV) on both approaches in the context of a case study predicting the incidence of COVID‐19 based on wastewater data. The TSCV framework outlined in the paper begins by engineering interpretable features hypothesized as potential predictors of COVID‐19 incidence. Feature selection and hyperparameter tuning are then utilized with TSCV to identify the best features and hyperparameters for optimal model performance given a specific forecast horizon. While evidence supporting the utility of TSCV for auto‐regressive integrated moving average model with exogenous variables (TS‐ARIMAX) forecasts is lacking in this study, such an approach proves advantageous for gradient boosting machine forecasts (TS‐GBM). In Wyoming, for instance, TS‐GBM had a 34.9% improvement compared to naïve predictions, whereas GBM without TSCV only had a 15.6% improvement. However, TSCV also enhances interpretability for both TS‐ARIMAX and TS‐GBM models as this approach selects specific features, such as lagged values of COVID‐19 cases, based on forecast performance and forecast length. Future research should work to explore the influence of stationarity and model averaging on the performance of TSCV in forecasting applications.

Application of multi-modal temporal neural network based on enhanced sparrow optimization in lithium battery life prediction

This paper introduces the DeNet-Mamba-DC-SCSSA network, an advanced solution for predicting the Remaining Useful Life (RUL) of lithium-ion batteries, crucial for the safety and efficiency management of electric vehicles. Combining the robust Denoising Enhancement Network (DeNet), the Improved Sparrow Optimization Algorithm (SCSSA), the adept Mamba time-series model, and the proficient Dilated Convolution (DC), this model excels in precise noise handling and sophisticated feature extraction. DeNet diligently refines input data, mitigating noise interference, while Mamba skillfully captures sequential intricacies. DC, on the other hand, adeptly extracts features over varying time scales, ensuring meticulous RUL predictions.The model’s efficacy was rigorously tested on NASA and CALCE datasets and was benchmarked against cutting-edge algorithms. Remarkably, it reduced average RE and RMSE by 48.59% and 21.45%, respectively, showcasing its superior performance and accuracy. Further evaluation on the CALCE dataset against the latest methods affirmed its leading predictive precision and stability.The model’s robustness and practical applicability were further validated using real vehicle data from a new energy vehicle platform. In a challenging test, it accurately predicted the charging capacities corresponding to the mileage of four vehicles with minimal errors: 0.52 Ah, 1.03 Ah, 0.84 Ah, and 0.71 Ah. These results significantly surpassed those of other recent methods, highlighting the model’s exceptional generalizability and potential for real-world applications in electric vehicle battery management.

Research on stock prediction based on CED-PSO-StockNet time series model

To tackle the challenge of low accuracy in stock prediction within high-noise environments, this paper innovatively introduces the CED-PSO-StockNet time series model. Initially, the model decomposes raw stock data using the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) technique and reconstructs the components by estimating their frequencies via the extreme point method. This process enhances component stability and mitigates noise interference. Subsequently, an Encoder-Decoder framework equipped with an attention mechanism is employed for precise prediction of the reconstructed components, facilitating more effective extraction and utilization of data features. Furthermore, this paper utilizes an Improved Particle Swarm Optimization (IPSO) algorithm to optimize the model parameters. On the Pudong Bank dataset, through ablation experiments and comparisons with baseline models, various optimization strategies incorporated into the proposed CED-PSO-StockNet model were effectively validated. Compared to the standalone LSTM model, CED-PSO-StockNet achieved a remarkable 45.59% improvement in the R2 metric. To further assess the model’s generalization capability, this paper also conducted comparative experiments on the Ping An Bank dataset, and the results underscored the significant advantages of CED-PSO-StockNet in the domain of stock prediction.

Study On Mathematical Model For Forecasting of Novel Coronavirus in India

This paper reveals that the new coronavirus (Covid-19) is the biggest challenge for the whole world. The World Health Organization (WHO) has declared it an epidemic. The data was collected from 209 different individual WHO situation reports for COVID-19 in India. First, prognostic models were compared based on minimum AIC, MAPE, and MAE. Then the best forecasting models were used on the epidemiologic data of India to predict the epidemiologic pattern of prevalence. New and Total Covid-19 Deaths and Incidence ARIMA and SARIMA have been found suitable and predicted for September 1, 2020. Also, the prophet model is used for accuracy and prediction of the total number of coronavirus cases in India. The predicted values are checked against the past observed values so that the two values are very close. Using such time series models we can forecast for the next 15-20 days and plan accordingly. This kind of projection helps in planning for the future.

Financial Sector Reforms and Inflation-Growth Nexus

In the 1980s, Pakistan's financial reforms reduced central bank borrowing, raising government debt and tax collection needs. In this context, the conventional approach to evaluating the costs and benefits of low inflation proves inadequate. The sacrifice ratio (SR) focuses on demand contraction's impact on economic activity and fails to capture the true costs of inflation by overlooking the supply-side effects of reduced fiscal space. This study re-evaluates the SR, accounting for both the demand contraction channel and the long-term supply side impact of diminished fiscal space. We hypothesize that the cumulative costs outweigh the long-term benefits of price stability. By incorporating fiscal space scenarios into the time-series model for SR estimation; our results indicate an SR approximately double that found in earlier studies. These findings highlight the importance of incorporating fiscal implications into monetary policy decisions to support sustainable economic growth.

Dynamic pricing and demand forecasting: Integrating time-series analysis, regression models, machine learning, and competitive analysis

Abstract. Dynamic pricing is a critical strategy for businesses seeking to optimize revenue and stay competitive in fluctuating markets. This paper explores the integration of various demand forecasting techniques, including time-series analysis, regression models, and machine learning algorithms, with competitive analysis methodologies to enhance dynamic pricing strategies. Time-series analysis focuses on decomposing data into trend, seasonality, and random fluctuations, using ARIMA models for accurate demand prediction. Regression models delve into the complexities of variable interactions, extending beyond linear relationships to include advanced techniques like Ridge, Lasso, and Elastic Net regression. Machine learning algorithms, such as decision trees, random forests, gradient boosting, and neural networks, revolutionize demand forecasting by uncovering complex patterns in large datasets. Competitive analysis incorporates market scanning, price elasticity estimation, and competitor behavior modeling to inform dynamic pricing decisions. Optimization algorithms, including linear programming, genetic algorithms, and simulated annealing, are employed to refine pricing strategies, while revenue management techniques like yield management and overbooking ensure maximum revenue from perishable goods and services. This comprehensive approach enables businesses to dynamically adjust prices in real-time, maintaining competitiveness and maximizing profitability in an ever-evolving market landscape.

Self-attention-based Diffusion Model for Time-series Imputation

Time-series modeling is essential for applications in agriculture, weather forecasting, food production, and more. However, missing data due to sensor malfunctions, power outages, and human errors is a common issue, complicating the training of machine learning models. We propose a diffusion-based generative model to address this problem and fill the gaps in the data. Our approach captures feature and time correlations through a two-stage imputation process. Our model outperforms state-of-the-art imputation methods and is more scalable in GPU resources.

Maximum-likelihood estimation of the Po-MDDRCINAR(p) model with analysis of a COVID-19 data

Integer-valued data are frequently encountered in time series studies. A pth-order mixed dependence-driven random coefficient integer-valued autoregressive time series model (Po-MDDRCINAR(p)) in view of binomial and negative binomial operators, where the innovation sequence follows a Poisson distribution, is investigated to provide meaningful theoretical explanations. Strict stationary and ergodicity of the model are demonstrated. Furthermore, the conditional least-squares and conditional maximum-likelihood methods are adopted to estimate the parameters, where the asymptotic characterization of the estimators is derived. Finite-sample properties of the conditional maximum-likelihood estimator are examined in relation to the widely used conditional least-squares estimator. The conclusion is that, if the Poisson assumption of the innovation sequence can be justified, conditional maximum-likelihood method performs better in terms of MADE and MSE. Finally, the practical performance of the model is illustrated by a set of COVID-19 data of suspected cases in China with a comparison with relevant models that exist so far in the literature.

Energy Market Price Forecasting and Financial Technology Risk Management Based on Generative AI

Abstract: The volatility of global energy markets, particularly electricity prices, plays a crucial role in influencing international economic activities. With the ongoing global energy transition and the push for low-carbon development, predicting electricity prices has become increasingly important for policymakers and market participants. This paper explores the forecasting capabilities of the ARIMA and LSTM models in analyzing electricity prices in the United States, drawing from data spanning 2001 to 2024.ARIMA, a traditional time series model, is valued for its simplicity and effectiveness in capturing linear trends, while LSTM, a deep learning-based model, excels at handling long-term dependencies in complex datasets. This study reveals that while both models offer valuable insights, each exhibits limitations. ARIMA struggles with non-linear patterns and volatility, whereas LSTM tends to underestimate extreme price values. The findings highlight the potential of hybrid models that combine traditional and machine learning approaches to enhance forecasting accuracy in the increasingly dynamic energy market. This research provides essential guidance for improving decision-making processes in the context of the global shift towards clean energy.

Selecting a Time-Series Model to Predict Drinking Water Extraction in a Semi-Arid Region in Chihuahua, Mexico

As the effects of global climate change intensify, it is increasingly important to implement more effective water management practices, particularly in arid and semi-arid regions such as Meoqui, Chihuahua, situated in the arid northern center of Mexico. The objective of this study was to identify the optimal time-series model for analyzing the pattern of water extraction volumes and predicting a one-year forecast. It was hypothesized that the volume of water extracted over time could be explained by a statistical time-series model, with the objective of predicting future trends. To achieve this objective, three time-series models were evaluated. To assess the pattern of groundwater extraction, three time-series models were employed: the seasonal autoregressive integrated moving average (SARIMA), Prophet, and Prophet with extreme gradient boosting (XGBoost). The mean extraction volume for the entire period was 50,935 ± 47,540 m3, with a total of 67,233,578 m3 extracted from all wells. The greatest volume of water extracted has historically been from urban wells, with an average extraction of 55,720 ± 48,865 m3 and a total of 63,520,284 m3. The mean extraction volume for raw water wells was determined to be 20,629 ± 19,767 m3, with a total extraction volume of 3,713,294 m3. The SARIMA(1,1,1)(1,0,0)12 model was identified as the optimal time-series model for general extraction, while a “white noise” model, an ARIMA(0,1,0) for raw water, and an SARIMA(2,1,1)(2,0,0)12 model were identified as optimal for urban wells. These findings serve to reinforce the efficacy of the SARIMA model in forecasting and provide a basis for water resource managers in the region to develop policies that promote sustainable water management.