Autoregressive Integrated Moving Average Time Series Research Articles

Management of Exchange Rate Forecasting Through Vanilla Long Short-Term Memory Networks (LSTM) and Auto-Regressive Integrated Moving Average (ARIMA)

Predicting foreign exchange rates presents a formidable challenge within financial forecasting, given its pivotal role in influencing a country's economic trajectory. To address this challenge, numerous forecasting models are employed with the aim of anticipating future exchange rate movements. This study aims to determine the efficacy of two prominent machine learning models, namely Vanilla Long Short-Term Memory Networks (LSTM) and Auto-Regressive Integrated Moving Average (ARIMA), in forecasting the exchange rate between the Malaysian Ringgit (MYR) and the United States Dollar (USD). Employing Python's robust statistical packages for time series forecasting, both Vanilla LSTM and ARIMA models undergo rigorous training on the dataset. Leveraging Python's programming capabilities enables in-depth analysis, essential for model refinement and accuracy assessment. Upon comparing the error measures of both models, it becomes evident that the Vanilla LSTM model outperforms ARIMA, exhibiting lower Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) values. Specifically, the MSE and RMSE for Vanilla LSTM stand at 0.0102 and 0.1011, respectively, surpassing ARIMA's 0.0113 and 0.1062. Thus, affirming Vanilla LSTM's superiority in exchange rate forecasting. Consequently, the study concludes that Vanilla LSTM emerges as the most accurate model for exchange rate prediction, with a projected exchange rate of RM4.22 for July 2022.

Analysis of the Impact of the “Insurance & Futures” Model on the Price Fluctuation of China's Pig Market

The market price cyclical fluctuation and oscillation "pig cycle" is a typical economic phenomenon in the pig market. In this paper, we use the coefficient of variation method and ARIMA (Auto-Regressive, Integrated, Moving Average) time-series method to analyse the weekly data of local breed pigs, domestic hybrid pigs, exotic hybrid pigs and hog price indexes from 2019 to 2023, and empirically analyse the impact of the "insurance + futures" model on the price of hogs to empirically analyse the degree of influence of the "insurance+futures" model on the volatility of hog prices. It is found that the price change trends of local breed pigs, domestic hybrid pigs, and exotic hybrid pigs are similar, and the "insurance + futures" model has a significant impact on the price fluctuation of exotic hybrid pigs, a weak impact on the price fluctuation of local breed pigs, and a non-significant impact on the price fluctuation of domestic hybrid pigs. During the pilot practice of the "insurance + futures" model, the price volatility of the three types of pigs increased significantly, and the overall pig price showed a downward trend.

Leveraging Large Data, Statistics, and Machine Learning to Predict the Emergence of Resistant E. coli Infections.

Drug-resistant Gram-negative bacterial infections, on average, increase the length of stay (LOS) in U.S. hospitals by 5 days, translating to approximately $15,000 per patient. We used statistical and machine-learning models to explore the relationship between antibiotic usage and antibiotic resistance over time and to predict the clinical and financial costs associated with resistant E. coli infections. We acquired data on antibiotic utilization and the resistance/sensitivity of 4776 microbial cultures at a Kaiser Permanente facility from April 2013 to December 2019. The ARIMA (autoregressive integrated moving average), neural networks, and random forest time series algorithms were employed to model antibiotic resistance trends. The models' performance was evaluated using mean absolute error (MAE) and root mean squared error (RMSE). The best performing model was then used to predict antibiotic resistance rates for the year 2020. The ARIMA model with cefazolin, followed by the one with cephalexin, provided the lowest RMSE and MAE values without signs of overfitting across training and test datasets. The study showed that reducing cefazolin usage could decrease the rate of resistant E. coli infections. Although piperacillin/tazobactam did not perform as well as cefazolin in our time series models, it performed reasonably well and, due to its broad spectrum, might be a practical target for interventions in antimicrobial stewardship programs (ASPs), at least for this particular facility. While a more generalized model could be developed with data from multiple facilities, this study acts as a framework for ASP clinicians to adopt statistical and machine-learning approaches, using region-specific data to make effective interventions.

Wind speed prediction using LSTM and ARIMA time series analysis models: A case study of Gelibolu

Wind energy stands out as a prominent renewable energy source, characterized by its high efficiency, feasibility, and wide applicability. Nonetheless, the integration of wind energy into the electrical system encounters significant obstacles due to the unpredictability and variability of wind speed. Accurate wind speed prediction is essential for estimating the short-, medium-, and long-term power output of wind turbines. Various methodologies and models exist for wind speed time series prediction. This research paper proposes a combination of two approaches to enhance forecasting accuracy: deep learning, particularly Long Short-Term Memory (LSTM), and the Autoregressive Integrated Moving Average (ARIMA) model. LSTM, by retaining patterns over longer periods, improves prediction rates. Meanwhile, the ARIMA model enhances the likelihood of staying within predefined boundaries. The study utilizes daily average wind speed data from the Gelibolu district of Çanakkale province spanning 2014 to 2021. Evaluation using the root mean square error (RMSE) shows the superior forecast accuracy of the LSTM model compared to ARIMA. The LSTM model achieved an RMSE of 6.3% and a mean absolute error of 16.67%. These results indicate the potential utility of the proposed approach in wind speed forecasting, offering performance comparable to or exceeding other studies in the literature.

Aquila Optimizer-Based Hybrid Predictive Model for Traffic Congestion in an IoT-Enabled Smart City

Effective traffic congestion prediction is need of the hour in a modern smart city to save time and improve the quality of life for citizens. In this study, AB_AO (ARIMA Bi-LSTM using Aquila optimizer), a hybrid predictive model, is proposed using the most effective time-series data prediction statistical model ARIMA (Autoregressive Integrated Moving Average) and sequential predictive Deep Learning (DL) technique LSTM (Long Short-Term Memory) which helps in traffic congestion prediction with a minimum error rate. Also, the Aquila optimizer (AO) is used to elevate the adequacy of the AB_AO model. Three road traffic datasets of different cities from the “CityPulse EU FP7 project” are used to implement the proposed hybrid model. In a time-series dataset, two components need to be handled with care, i.e., linear and nonlinear. In this study, the ARIMA model has been used to manage linear components and Bi-LSTM is used to handle nonlinear components of the time-series dataset. The Aquila Optimizer (AO) is used for hyperparametric tuning to enhance the performance of Bi-LSTM. Error measurement parameters like the Mean Absolute Error (MAE), Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE) are used to validate the results. A detailed mathematical and empirical analysis is given to justify the performance of the AB_AO model using an ablation study and comparative analysis. The AB_AO model acquires more stable and precise results with MSE as 18.78, MAE as 3.18, and MAPE as 0.21 than other models. It may further help to predict the vehicle count on the road, which may be of great help in reducing wastage of time in traffic congestion.

Financial Statement Analysis Based on RNN-RBM Model

Financial statement analysis is a critical component of decision-making for businesses, investors, and financial professionals. To enhance the accuracy and effectiveness of such analysis, this paper introduces the application of an innovative approach known as the Intelligent Swarm Regression ARIMA Model. This advanced model combines the power of swarm intelligence with ARIMA (AutoRegressive Integrated Moving Average) time series forecasting, offering a robust methodology for predicting and analyzing key financial metrics. The study begins by providing an overview of the Intelligent Swarm Regression ARIMA model and its application to financial data. Through a comprehensive analysis of financial statements, including market capitalization, revenue, net income, and other crucial indicators, the model's efficacy in predicting future values is evaluated. Additionally, the paper examines the deviations between predicted and actual financial values, offering insights into the model's accuracy and areas for potential improvement. The findings of this research are invaluable for investors, financial analysts, and companies seeking to optimize their financial performance and strategic decision-making. By leveraging the Intelligent Swarm Regression ARIMA Model, stakeholders can make well-informed choices that lead to better financial outcomes and a competitive advantage in a dynamic economic landscape. This paper represents a significant step forward in the financial analysis, providing a practical methodology and a pathway to enhanced financial decision-making. As the importance of financial data continues to grow, this research offers a promising avenue for achieving financial success and stability.

Analyzing the Resilience of Amazon and Darden Amidst the COVID-19 Pandemic: A Time-Series Study with ARIMA Modeling

On March 16, 2020, in response to the emergence of the COVID-19 virus, the United States implemented an indefinite quarantine mandate. This unprecedented measure had repercussions, affecting businesses, schools, and government facilities across the nation. The mandate required these entities to either operate at reduced capacity or to halt operations altogether. Consequently, the stock market experienced extreme volatility during this period. This study aims to conduct an in-depth analysis of the responses to the pandemic of two firms operating in different industries. Amazon, an e-commerce and technology giant in the broadband retail industry, found itself with unprecedented demand for online shopping and cloud services considering the prevalent supply chain disruptions at the time. On the flip side, Darden Restaurants, a prominent player in the restaurant industry, had to either operate at a limited capacity, or shut down most of its restaurants altogether in coordination with the mandated closures issued by the government. This study will employ time series analysis techniques, specifically Autoregressive (AR), Moving Average (MA), and ARIMA (Autoregressive Integrated Moving Average) models. The objective is to assess the influence of COVID-19 on the selected firms and extract valuable insights from the data.

Design of a Framework for Demand Forecasting Using Time Series Decomposition-Based Approach

In recent years, artificial intelligence (AI) has made highly accurate demand forecasting possible. However, improving forecast accuracy does not necessarily mean reducing inventory costs or improving service levels in supply chain and inventory management, which are closely related to demand forecasting. Workers require not only high accuracy but also a basis for making decisions a based on forecasts. Autoregressive integrated moving average (ARIMA) and seasonal autoregressive integrated moving average (SARIMA) are demand forecasting methods with high accuracy and interpretability. However, these methods cannot provide evidence for demand fluctuations such as trends and seasonality, although they exhibit an autoregressive time-series structure. In this study, a framework for demand forecasting with high accuracy and interpretability was designed using time series decomposition and ARIMA to support decision makers in demand forecasting. The Seasonal-trend decomposition using locally estimated scatterplot smoothing (STL) is used to decompose a time series into three components trend, seasonality, and residual to provide decision makers with an easily understandable basis for demand changes. In addition, the ARIMA model is used for trends and residuals to achieve highly accurate forecasts. Comparing the prediction accuracies of the proposed STL-ARIMA and SARIMA models shows that STL-ARIMA has higher interpretability than the SARIMA model and the same accuracy.

A novel framework for machine intelligence aided smarter energy management in residential scenario

Energy management systems can monitor, optimize, and control energy utilization in residential and commercial applications. Different solutions are available for achieving better energy management. Among the different solutions, load forecasting is considered as one of most preferred technique. It is essential for a load forecasting technique to predict the aggregated loads at the residential level. This paper presents a systematic analysis of different load forecasting models considering a hybrid renewable energy based smart home which uses a solar photovoltaic (PV) system, wind system, and battery storage system for energy storage. In this work, the various forecasting techniques such as time-series models, regression models, ARIMA (Autoregressive Integrated Moving Average) models, and neural network models are analyzed for different load cases. Simulation is performed based on the effect of seasonal bias and condition load shifting on energy generation. The obtained results show that both polynomial and SVM regression model achieves better performance in terms of R2 score and adjusted R2 score of 0.73, and RMSE values of 0.84. It can also be inferred from the analysis that, the random forest regression model and a neural network-based-long short-term memory (LSTM) model exhibit superior results in comparison to other models.

Sensitivity analysis for forecasting Brazilian electricity demand using artificial neural networks and hybrid models based on Autoregressive Integrated Moving Average

Several studies focus on improving forecasting techniques to capture multiple patterns in time series. The evolution of computing hardware has made possible to solve complex equations with large amount of data, such as the one used in neural networks. On the other hand, time series methods such as ARIMA (Autoregressive Integrated Moving Average) could also have a good approximation with low computational resources. Nonetheless, to improve the ARIMA approximations, it could be combined with other techniques such as Wavelet Transform or Fourier Transform. Therefore, this work evaluates the appropriate utilization to make predictions for different time horizons (2, 5 and 10 years) and different time frequencies (days, months, and years) using artificial neural network, ARIMA combined with Wavelet Transform, or Fourier Transform. The results show that Artificial Neural Networks provides a better approach for short-term horizons considering either time frequency, ARIMA with Fourier Transform has the best approximation for the monthly time series and either time horizons and ARIMA with Wavelet Transform has the best approximation for medium-term and long-term periods with either time frequency.

Analysis of Amazon Stock Using Simple Linear Regression and Time Series ARIMA Model

The rate of daily return of a stock is one of the important indicators for investors to anticipate benefits or losses from historical data. This paper will focus on the stock of Amazon, which is a popular choice for stock traders and contains data from August 25, 2017, to August 24, 2022. By using regression models such as a simple linear regression model and an autoregressive integrated moving average (ARIMA), the recent daily return value is predicted based on data during these 5 years. The simple linear regression can show the trend of stock price and the predicted response rate of daily return using the linear model. Furthermore, ARIMA is a more advanced time series model to provide a more accurate rate of daily return with confidence intervals. The predicted trend and rate of daily return are useful for investors to make decisions to buy or sell a stock recently. The trend can tell investors whether the stock price would go up and the daily return can indicate how many benefits can they earn if they choose to invest in this stock.

Using weather factors and google data to predict COVID-19 transmission in Melbourne, Australia: A time-series predictive model

BackgroundForecast models have been essential in understanding COVID-19 transmission and guiding public health responses throughout the pandemic. This study aims to assess the effect of weather variability and Google data on COVID-19 transmission and develop multivariable time series AutoRegressive Integrated Moving Average (ARIMA) models for improving traditional predictive modelling for informing public health policy. MethodsCOVID-19 case notifications, meteorological factors and Google data were collected over the B.1.617.2 (Delta) outbreak in Melbourne, Australia from August to November 2021. Timeseries cross-correlation (TSCC) was used to evaluate the temporal correlation between weather factors, Google search trends, Google Mobility data and COVID-19 transmission. Multivariable time series ARIMA models were fitted to forecast COVID-19 incidence and Effective Reproductive Number (Reff) in the Greater Melbourne region. Five models were fitted to compare and validate predictive models using moving three-day ahead forecasts to test the predictive accuracy for both COVID-19 incidence and Reff over the Melbourne Delta outbreak. ResultsCase-only ARIMA model resulted in an R squared (R2) value of 0.942, Root Mean Square Error (RMSE) of 141.59, and Mean Absolute Percentage Error (MAPE) of 23.19. The model including transit station mobility (TSM) and maximum temperature (Tmax) had greater predictive accuracy with R2 0.948, RMSE 137.57, and MAPE 21.26. ConclusionMultivariable ARIMA modelling for COVID-19 cases and Reff was useful for predicting epidemic growth, with higher predictive accuracy for models including TSM and Tmax. These results suggest that TSM and Tmax would be useful for further exploration for developing weather-informed early warning models for future COVID-19 outbreaks with potential application for the inclusion of weather and Google data with disease surveillance in developing effective early warning systems for informing public health policy and epidemic response.

Time Series Analysis of Monthly Coffee (Robusta) Prices in India using Box-Jenkins Approach

Robusta coffee is a type of coffee made from the Coffea canephora plant's beans (seeds). It is the world's second most popular coffee, accounting for 43% of global coffee production with arabica constituting the remainder except for the 1.5% constituted by Coffea liberica. The purpose of this study is to focus on predicting monthly coffee prices in India by using the historic time series data. The objective of this paper is to fit an Autoregressive Integrated Moving Average model using Box-Jenkins approach. Numerous fields, including agricultural production, animal husbandry and dairy economics, stock price prediction, etc. depend heavily on forecasting. To choose the best model, Autoregressive (AR), Moving Average (MA), and Autoregressive Integrated Moving Average (ARIMA) processes was used to select the best model for monthly coffee prices in India. This paper discusses ARIMA (p, d, q) time series analysis and its components, ACF, PACF, Normalized BIC, Box-Ljung Q Statistics, and Residual analysis. According to the best fitted model i.e., ARIMA (0,2,1) monthly coffee prices in India is expected to increase to INR 89.35 kg-1 in the month of November 2022. The outcomes are represented numerically and graphically.

Multiple crop yield estimation and forecasting using MERRA-2 model, satellite-gauge and MODIS satellite data by time series and regression modelling approach

Multiple crop yield estimation and forecasting have become essential because climate change and land degradation are deteriorating production of crops. The present study presents multiple crop yield estimation and forecasting efficiently and precisely using MERRA-2 model, satellite-gauge and MODIS-Terra satellite data by regression and time series models. Rice, wheat, maize, rapeseed & mustard, and millet crops yield data of India was used for crop yield estimation and forecasting using Multiple Linear Regression (MLR), Decomposition, Holt, Winters and Autoregressive Integrated Moving Average (ARIMA) models. Models were developed using crops yield historical data, weather variables (temperature, precipitation, surface radiation) and pollutants (O3, CO) data of 1982–83 to 2014–15. Overall best correlation (R = 0.96) and poor correlation (R = 0.56) was found between observed and estimated rice and rapeseed & mustard crops using Decomposition and MLR models, respectively. After rigorous skill assessment and out of more than 20 combinations of ARIMA models, the ARIMA (0, 1, 1) time series model performed the best with Root Mean Square Error (RMSE = 4.59%) for wheat crop estimation. Forecasted crop yields were successfully cross-validated using four years’ data (2015–16, 2016–17, 2017–18 and 2018–19) for regression and time series models by comparing Relative Error (RE%) between observed and forecasted crops yields. Results were statistically analyzed using correlation coefficient (R), Bayesian Information Criteria (BIC), RMSE%, Mean Absolute Percentage Error (MAPE%), t-value and p-value. The crop yield estimation and forecasting results showed reasonably quantifiable information with the actual datasets. The outcomes of the present study have policy level implications for compensation to farming communities for the loss of crops.

Impact of a dental care intervention on the hospital mortality of critically ill patients admitted to intensive care units: A quasi-experimental study

We aimed to evaluate the impact of providing dental care to critically ill patients on their risk of death and ventilator-associated pneumonia (VAP). A quasi-experimental study was conducted in 2 intensive care units (ICU) from 2016 to 2019. The intervention consisted of implementing routine dental care, focusing on oral hygiene and periodontal treatment, at least 3 times a week, for patients admitted to the study units. In the pre-intervention period, routine oral hygiene was provided by the ICU nursing staff. The primary and secondary study outcomes were mortality, evaluated at the end of the ICU stay, and VAP incidence density, respectively. Data were analyzed using the ARIMA (autoregressive integrated moving average) time series model in R software. During the intervention period, 5,147 dental procedures were performed among 355 patients. The time series showed that ICU mortality was 36.11%, 32.71%, and 32.30% within the 3 years before the intervention, and 28.71% during the intervention period (P=.015). VAP incidence density did not significantly change during the study period (P=.716). A dental care intervention focused on oral hygiene and periodontal treatment regularly provided by dentists to critically ill patients may decrease their risk of dying in the ICU. Randomized clinical trials should be performed to confirm these findings. WHO-affiliated Brazilian Clinical Trials Registry. RBR-4jmz36. Registered 7 October 2018, before first patient enrollment.

Back-Propagation Neural Network and ARIMA Algorithm for GDP Trend Analysis

GDP (gross domestic product) is a key indicator for assessing a country’s or region’s macroeconomic situation, as well as a foundation for the government to develop economic development strategies and macroeconomic policies. Currently, the majority of methods for forecasting GDP are linear methods, which only take into account the linear factors that affect GDP. GDP (gross domestic product) is widely regarded as the most accurate indicator of a country’s economic health. GDP not only reflects a country’s economic development over time but can also reflect its national strength and wealth. As a result, the GDP trend forecast partially reflects China’s transformation and future development. The time series ARIMA (Autoregressive Integrated Moving Average) model and the BPNN (BP neural network) model are combined in this article to create the ARIMA-BPNN fusion prediction model. The predicted values of the two models were then weighted averaged to obtain the predicted values of the linear part of the improved fusion model. To get the predicted values of the improved fusion model, we weighted average the residual parts of the two models, predict the nonlinear residual with BPNN, and add the predicted values of the two parts. It is applied to the actual GDP forecast in H province from 2019 to 2022, and the actual forecast verifies the effectiveness of the fusion forecast model in the actual forecast.

Multi-Regional Modeling of Cumulative COVID-19 Cases Integrated with Environmental Forest Knowledge Estimation: A Deep Learning Ensemble Approach.

Reliable modeling of novel commutative cases of COVID-19 (CCC) is essential for determining hospitalization needs and providing the benchmark for health-related policies. The current study proposes multi-regional modeling of CCC cases for the first scenario using autoregressive integrated moving average (ARIMA) based on automatic routines (AUTOARIMA), ARIMA with maximum likelihood (ARIMAML), and ARIMA with generalized least squares method (ARIMAGLS) and ensembled (ARIMAML-ARIMAGLS). Subsequently, different deep learning (DL) models viz: long short-term memory (LSTM), random forest (RF), and ensemble learning (EML) were applied to the second scenario to predict the effect of forest knowledge (FK) during the COVID-19 pandemic. For this purpose, augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) unit root tests, autocorrelation function (ACF), partial autocorrelation function (PACF), Schwarz information criterion (SIC), and residual diagnostics were considered in determining the best ARIMA model for cumulative COVID-19 cases (CCC) across multi-region countries. Seven different performance criteria were used to evaluate the accuracy of the models. The obtained results justified both types of ARIMA model, with ARIMAGLS and ensemble ARIMA demonstrating superiority to the other models. Among the DL models analyzed, LSTM-M1 emerged as the best and most reliable estimation model, with both RF and LSTM attaining more than 80% prediction accuracy. While the EML of the DL proved merit with 96% accuracy. The outcomes of the two scenarios indicate the superiority of ARIMA time series and DL models in further decision making for FK.

Determination ARIMA Model for the Discharge Acid Concentration in Pressure Oxidation Autoclave at A Gold Processing Plant

During industrial production, many quality parameters of products are measured to evaluate the quality of final products. These investigations are also applied for the machine parameters to control if they work properly. A huge amount of data is collected, recorded, and evaluated for these aims. Time series analyses are often a useful tool to evaluate the industrial data to analyze the products quality parameters or machine performances. The autoregressive integrated moving average (ARIMA) time series models are the most known and used method. One of the most important advantages of ARIMA models is its capability of near future estimation for the monitored process variable. For this aim, current research is carried out at a gold processing plant data. The plant gains the gold by applying the modern pressure oxide leaching in autoclave and cyanide leaching. Since the gold ore characteristics can be changed even, they are mined in the same mining area, monitoring the product quality and machine performances during production stages are needed. In this research, the data set, which was recorded at equal sampling time intervals and obtained in a month, is supplied from a gold processing plant in Turkey. During the gold production, the ground gold ore is subjected to the desulphurization process by pressure oxide (POX) in the autoclave. Discharge acid concentration is critical in this process and it is followed regularly by taking samples in 2 hours’ time intervals in 24 hours. Using discharge acid concentration data set, the ARIMA (1,0,1) time series model was determined to monitor it. Also, near-future values of acid concentration were estimated by the ARIMA (1,0,1) model and compared with the real discharge acid concentrations. It was determined that there was a very good agreement between the estimated values obtained by the ARIMA (1,0,1) model and real values.

Towards optimal ATM cash replenishment using time series analysis

Automatic Teller Machine (ATM) are still largely used to dispense cash to the customers. ATM cash replenishment is a process of refilling ATM machine with a specific amount of cash. Due to vacillating users demands and seasonal patterns, it is a very challenging problem for the financial institutions to keep the optimal amount of cash for each ATM. In this paper, we present a time series model based on Auto Regressive Integrated Moving Average (ARIMA) technique called Time Series ARIMA Model for ATM (TASM4ATM). This study used ATM back-end refilling historical data from 6 different financial organizations in Pakistan. There are 2040 distinct ATMs and 18 month of replenishment data from these ATMs are used to train the proposed model. The model is compared with the state-of- the-art models like Recurrent Neural Network (RNN) and Amazon’s DeepAR model. Two approaches are used for forecasting (i) Single ATM and (ii) clusters of ATMs (In which ATMs are clustered with similar cash-demands). The Mean Absolute Percentage Error (MAPE) and Symmetric Mean Absolute Percentage Error (SMAPE) are used to evaluate the models. The suggested model produces far better forecasting as compared to the models in comparison and produced an average of 7.86/7.99 values for MAPE/SMAPE errors on individual ATMs and average of 6.57/6.64 values for MAPE/SMAPE errors on clusters of ATMs.

Forecasting the seasonal natural gas consumption in the US using a gray model with dummy variables

To forecast the seasonal fluctuations of US natural gas consumption accurately, a novel gray model based on seasonal dummy variables and its derived model are separately established. Then, the approximate time response formula in the proposed seasonal forecasting model is optimally calculated by using particle swarm optimization algorithm. On this basis, empirical analysis is conducted using data pertaining to natural gas consumption in the US during 2010–2019. The results show that gray model based on seasonal dummy variables and its derived model can recognize seasonal fluctuations in US natural gas consumption, whose prediction performances are much better than that of a traditional gray model, autoregressive integrated moving average (ARIMA), support vector regression (SVR), recurrent neural network (RNN), transformer, and fuzzy time series forecasting models. The mean absolute percentage errors (MAPEs) of the proposed seasonal gray forecasting model and its derived model, the classic gray model, ARIMA, SVR, RNN, Transformer, and Fuzzy time series forecasting models are 3.46%, 2.37%, 12.23%, 3.39%, 2.38%, 3.08%, 3.84%, and 5.02% in the training set, while those are 4.57%, 4.42%, 12.44%, 7.9%, 8.09%, 11.33%, 35.19% and 13.19% in the test set, respectively. The predicted and empirical results obtained by utilizing the proposed gray model implied that natural gas consumption in the US from 2020 to 2022 will maintain its seasonal growth and periodic changes, with the highest and the lowest values in the first and second quarters, respectively.