Mean Absolute Percentage Error Values Research Articles

State of charge estimation for lithium‐ion battery pack based on real vehicle data and optimized backpropagation method by adaptive cross mutation sparrow search algorithm

AbstractIn response to the issues of traditional backpropagation (BP) neural networks in state of charge (SOC) estimation, including easy convergence to local optima, slow convergence speed, and low accuracy, this paper proposes a novel adaptive crossover mutation strategy and dynamic sparrow search algorithm to optimize BP networks' initial values and thresholds (ACMSSA‐BP). The proposed method is based on the sparrow search algorithm, where the number of producers and scroungers is adjusted through an adaptive factor. This improvement effectively transitions the search process from extensive full exploration to localized fine‐tuning search. In the position update phase of the producers, crossover mutation and dynamic search strategies are introduced to increase the diversity of good populations, prevent the algorithm from converging to local optima, and maintain its local search capability in the later stage. Using real transportation data from coal mining flame‐proof tracked vehicles, we applied correlation theory to extract model feature parameters and constructed a training data set to estimate the SOC. The results of both static and dynamic validation experiments have indicated that the ACMSSA‐BP method has delivered impressive performance in predicting SOC, as reflected in the mean absolute error, root mean squared error, and mean absolute percentage error values of less than 1.5%, 1.5%, and 1.6%, respectively. Compared with BP, SSA‐BP, CMSSA‐BP, PSO‐BP, and NARX_NN methods, the ACMSSA‐BP approach demonstrates enhanced accuracy in estimation, significant robustness, and impressive generalization capabilities.

Estimation of greenhouse gas emissions using linear and logarithmic models: A scenario-based approach for Turkiye's 2030 vision

Turkiye pledged to considerably reduce its greenhouse gas (GHG) emissions by the year 2030 as a part of its commitment under the Paris Agreement. The problem with analyzing mitigation of greenhouse gas emissions requires generation of accurate, reliable and consistent emission forecasts. This study aimed to accurately forecast Turkiye's total CO2 emissions per capita and per capita CO2 emissions from energy industries, industrial processes and agricultural sectors till 2030 using linear and logarithmic models based on increasing and decreasing scenarios. Turkiye's total CO2 emissions per capita in 2030 could reach to 7.6 and 7.7 tons of CO2, with total emissions of about 635 and 643 million tons (Mt) based on linear and logarithmic models, respectively. Linear modeling results showed that per capita CO2 emissions from Turkiye's energy industries, industrial processes and agricultural sectors could reach to 5.3, 0.9 and 0.9 tons in 2030, respectively, while logarithmic modeling results yielded that per capita CO2 emissions from Turkiye's energy industries, industrial processes and agricultural sectors could be 5.5, 1.1 and 0.9 tons in 2030, respectively. The accuracy of fit for linear and logarithmic models was assessed by calculating root mean square error (RMSE < 0.2036) and mean absolute percentage error (MAPE < 12.3347) values which showed that the models fitted well with the timeline data. In conclusion, Turkiye's greenhouse gas emissions can be reduced if utilization of renewable energy sources in the country's energy portfolio is increased, and the shares of energy intensive processes in the industrial and agricultural sectors are reduced.

Forecasting COVID-19 Daily Infected Cases in Sri Lanka by Holt-Winters Exponential Smoothing Method

The novel coronavirus disease (COVID-19) has spread from China since December 2019 and spread worldwide including Sri Lanka. The aim of this study was to forecast the daily infected cases of COVID-19 in Sri Lanka which in turn help administrators for effective management of the pandemic. The method used in this study was Holt-Winters three parameter with additive or multiplicative models. The daily infected cases in Sri Lanka during the period of 22nd January 2020 to 22nd December 2021 were obtained from the publicly available databases of Epidemiology Unit of Sri Lanka and World Health Organization. The pattern recognition of the daily infected cases was examined by time series plot and Auto Correlation Function (ACF). The model validation was performed by the Anderson Darling test which confirmed the normality of residuals (p &gt; 0.05) and ACF that confirmed the independence of residuals of the model. The forecasting ability of the model was assessed by the three measurements of errors; Mean Absolute Percentage Error (MAPE), Mean Absolute Deviation (MAD) and Mean Square Error (MSE). Holt-Winters additive and multiplicative model with α (level) 0.61, β (trend) 0.4 and γ (seasonal) 0.3 at a length of repeating behaviour of 3 days, had the least relative and absolute measurement of errors during the model fitting and verification. İn the multiplicative model, MAPE, MAD and MSE were 0.2847, 0.0187 and 0.0005 respectively. Similarly in the additive model, corresponding values of MAPE, MAD and MSE were 0.0207, 0.0187 and 0.0005. The fits and the forecast of these models followed a similar pattern of the actual daily infected cases concluding that the Holt-Winters model can be used to forecast the COVID-19 outbreak in Sri Lanka.

Prediction of local scour depth around bridge piers: modelling based on machine learning approaches

Scouring is a threat to the foundations of bridges, and it frequently results in bridge collapse. Recently, machine-learning approaches have become quite popular in solving hydraulic and hydrologic problems. The present study adopted adaptive neuro-fuzzy inference systems (ANFIS) and gene expression programming (GEP) to model bridge pier scour depth. A temporal scour depth relationship has been proposed based on ANFIS and GEP approaches. In this paper, 500 data sets have been used to model the temporal scour depth using ANFIS and GEP models, in which 80% (400) data sets for training and 20% (100) datasets for testing purposes are used to develop the model. The performance of the GEP model is validated by comparing the results of the developed ANFIS model and other existing empirical equations. It is found that the GEP-based scour depth prediction model is effective in training and validation. The present GEP model accurately predicts the scour depth with a mean absolute percentage error (MAPE) value of less than 12% and an R2 value greater than 0.85. Thus, the GEP model can be utilized to predict scour depth around the bridge pier for unsteady flow conditions.

Long-Term Prediction of Particulate Matter2.5 Concentration with Modal Autoformer Based on Fusion Modal Decomposition Algorithm

To overcome the limitations of long-term prediction of PM2.5 concentration, a multi-factor information flow causality analysis method is used to screen suitable meteorological and air pollutant-related factors and concatenate them with a PM2.5 sequence as the dataset. A modal decomposition algorithm is used as a module to be integrated into the autoformer (transformer improved with autocorrelation mechanism) model to improve it, and the modal autoformer (empirical modal decomposition combined with autoformer) is proposed. The constructed model decomposes the sequence into several components by using the modal decomposition module and uses the self-correlation mechanism and decomposition structure to decompose and extract features of different components at the time-feature level. Based on the matching method, the model is adjusted for different component features to improve the long-term prediction effect. The model is applied to three cities in Henan Province, Zhengzhou, Luoyang, and Zhumadian, as examples for experiments, and gated neural unit (GRU), informer, autoformer, and modal GRU (empirical modal decomposition combined with GRU model) are constructed for comparative verification. The results show that the modal autoformer can better cope with the complex characteristics of long-term prediction of the PM2.5 time series, has strong spatial adaptability and that its various indicators are optimal for the three cities, with R2 values being all above 0.96, where the highest is 0.987 in Zhengzhou; MAPE (Mean absolute percentage error) values all being less than 10, where the best is 7.602 in Zhumadian; and MAE (Mean absolute error) values all being less than 4. The prediction effect is stable enough, showing its feasibility and adaptability in long-term prediction.

Capacity Prediction Method of Lithium‐Ion Battery in Production Process Based on Improved Random Forest

Measuring capacity in the grading process is an important step in battery production. The traditional capacity acquisition method requires considerable time and energy consumption; therefore, an accurate capacity estimation is crucial in reducing production costs. Herein, a capacity prediction method for lithium‐ion batteries based on improved random forest (RF) is proposed. This method extracts features from the voltage data of the entire formation process and the first 25% of the grading process, saving 56.7% of the energy consumption and 74.6% of the time in the grading process. The importance of these features is ranked by RF, the best feature subset is selected, and the RF algorithm with parameters optimized by the genetic algorithm is applied to establish a capacity prediction model. The results show that the proposed model can accurately predict the capacity, with root mean square error (RMSE) and mean absolute percentage error (MAPE) values of only 191.89 mAh and 0.13%, respectively. Compared with other regression algorithms, the model shows a lower RMSE and MAPE and a higher capacity prediction accuracy for low‐capacity cells.

Selecting the accurate hydrological method for estimating peak discharge in the Lesti River Catchment Area, Malang Regency, East Java Province, Indonesia

Estimating peak discharge in the catchment poses a challenge for hydrologists due to the potential overflow risks. The present study examined various methods, including Synthetic unit hydrograph, Melchior, and Rational, to determine the peak discharge value in the Lesti River Catchment Area. Accurate hydrological input data is essential for effectively estimating peak discharge and mitigating potential damage or failure. Optimized physical parameters using geographic information systems are critical in generating peak discharge values, emphasizing their importance in the estimation process. The mean absolute percentage error (MAPE) and mean average error (MAE) were used to determine the accurate method for estimating peak discharge. Analysis results showed varied peak discharge values across different methods, and it found that the MAPE values ranged from 15.15% to 7.48% and the MAE values ranged from 1.47% to 0.63%, respectively. The Nakayasu synthetic unit hydrograph obtained the minimum error measure value compared to other methods by 7.48% in MAPE and 0.63% in MAE. Therefore, the MAPE and MAE performance considered that the Nakayasu was closely approximating the observed peak discharge and a relatively accurate method for estimating peak discharge in the Lesti River catchment area.

Predicting Axial Force in Friction Stir Welded Joints of AA7075-T6 and AA2024-T3 Using Machine Learning Techniques

The friction stir welding (FSW) process has gained popularity in joining various types of aluminum alloys. The heat generated during the process is caused by the friction between the shoulder of the welding tool and the workpiece. The axial force (Fz) of the weld plays a crucial role in the welding process. If the axial force is insufficient to generate heat will result in defective workpieces. Therefore, the aim of this study was to predict the axial force of the FSW process of dissimilar aluminum alloys (AA7075-T6 and AA2024-T3) using machine learning techniques. The data used to create the predictive model was obtained from experiments involving 5 factors, each with 2 levels: 1) welding speed (mm/min), 2) rotational speed (RPM), 3) type of tools, 4) plunge depth (mm), and 5) dwell time (sec). The axial force was measured using a dynamometer with a sampling frequency of 10 Hz. The predictive model was created using all 4 algorithms: AdaBoost, CatBoost, LightGBM, and XGBoost. The performance of the four predictive models was evaluated using four metrics: mean absolute error (MAE), mean absolute percent error (MAPE), mean square error (MSE), and root mean square error (RMSE). The results showed that the AdaBoost algorithm had the best performance, with MAE, MAPE, MSE, and RMSE values of 509.08, 0.24, 452591.73, and 672.75, respectively. The AdaBoost algorithm was then used to predict the axial force using a dataset of 29,282 data, with predicted minimum average axial force of 1871.6 N. When compared to the axial force measured in the experiment, which was 1265 N, the results showed that the AdaBoost algorithm was capable of predicting the axial force with acceptable accuracy.

On Hybrid Autoregressive Integrated Moving Average Artificial Neural Network Time Series Analysis of the Nigeria External Reserves

The amount of the external reserves possessed by a country is an essential component in the measure of its economic status. In essence, a country with substantial amount of money in its external reserves would experience a buoyant and friendly economy among other nations. This study employs three different time series models to examine the status of the Nigeria External reserves for 22 years (1960-2022). These include the Auto Regressive Integrated Moving Average (ARIMA), Artificial Neural Networks (ANN) and the hybrid ARIMA-ANN models. ARIMA (1,0,1)(0,0,2)12 emerged as the optimum model among other fitted linear models, having the smallest Akaike Information Criterion (AIC) value of 14365.84 and was used to fit the nonlinear Hybrid ARIMA-ANN model and for the prediction of the External Reserves series. The Mean Absolute Error (MAE), the Root Mean Square Error (RMSE) and the Mean Absolute Percentage Error (MAPE) values of (590.1479), (358.3421), and (24.10321) for the Hybrid ARIMA-ANN model were also the least in comparison with their corresponding values for the other models. These show that the Hybrid model has the least error value compared with those of the independent ARIMA and ANN models. Hence, the nonlinear Hybrid ARIMA-ANN model performed excellently in the estimation and the generation of the forecast values than the conventional linear ARIMA and the nonlinear ANN models.

Hybrid time series models with exogenous variable for improved yield forecasting of major Rabi crops in India

Accurate and in-time prediction of crop yield plays a crucial role in the planning, management, and decision-making processes within the agricultural sector. In this investigation, utilizing area under irrigation (%) as an exogenous variable, we have made an exertion to assess the suitability of different hybrid models such as ARIMAX (Autoregressive Integrated Moving Average with eXogenous Regressor)–TDNN (Time-Delay Neural Network), ARIMAX–NLSVR (Non-Linear Support Vector Regression), ARIMAX–WNN (Wavelet Neural Network), ARIMAX–CNN (Convolutional Neural Network), ARIMAX–RNN (Recurrent Neural Network) and ARIMAX–LSTM (Long Short Term Memory) as compared to their individual counterparts for yield forecasting of major Rabi crops in India. The accuracy of the ARIMA model has also been considered as a benchmark. Empirical outcomes reveal that the ARIMAX–LSTM hybrid modeling combination outperforms all other time series models in terms of root mean square error (RMSE) and mean absolute percentage error (MAPE) values. For these models, an average improvement of RMSE and MAPE values has been observed to be 10.41% and 12.28%, respectively over all other competing models and 15.83% and 18.42%, respectively over the benchmark ARIMA model. The incorporation of the area under irrigation (%) as an exogenous variable in the ARIMAX framework and the inbuilt capability of the LSTM model to process complex non-linear patterns have been observed to significantly enhance the accuracy of forecasting. The performance supremacy of other hybrid models over their individual counterparts has also been evident. The results also suggest avoiding any performance generalization of individual models for their hybrid structures.

Modelling Polyphenol Extraction through Ultrasound-Assisted Extraction by Machine Learning in Olea europaea Leaves.

The study of the phenolic compounds present in olive leaves (Olea europaea) is of great interest due to their health benefits. In this research, different machine learning algorithms such as RF, SVM, and ANN, with temperature, time, and volume as input variables, were developed to model the extract yield and the total phenolic content (TPC) from experimental data reported in the literature. In terms of extract yield, the neural network-based ANNZ-L model presents the lowest root mean square error (RMSE) value in the validation phase (9.44 mg/g DL), which corresponds with a mean absolute percentage error (MAPE) of 3.7%. On the other hand, the best model to determine the TPC value was the neural network-based model ANNR, with an RMSE of 0.89 mg GAE/g DL in the validation phase (MAPE of 2.9%). Both models obtain, for the test phase, MAPE values of 4.9 and 3.5%, respectively. This affirms that ANN models would be good modelling tools to determine the extract yield and TPC value of the ultrasound-assisted extraction (UAE) process of olive leaves under different temperatures, times, and solvents.

Ultimate axial strength prediction of concrete-filled double-skin steel tube columns using soft computing methods

In this study, two hybrid models, grey wolf optimizer (GWO) combined with group method of data handling (GMDH) and random forest (RF) optimized by particle swarm optimization (PSO), are proposed to predict the ultimate axial strength of concrete-filled double-skin steel tube (CFDST) columns. 139 sets of data collected from the literature were used to build the proposed models. The proposed hybrid models were compared with other machine learning models as well as empirical equations and design specifications. The results show that for all datasets, the integral absolute error (IAE), root relative squared error (RRSE), mean absolute percentage error (MAPE) and coefficient of determination (R2) predicted by PSO-RF and GWO-GMDH models are 0.057, 0.092, 0.068 and 0.992; 0.068, 0.094, 0.110 and 0.989, respectively. Compared with the single RF and GDMH models, the IAE, RRSE and MAPE values of the proposed PSO-RF and GWO-GMDH models decreased by 3.91%, 13.66% and 6.34%; 37.81%, 48.78% and 39.24%, respectively, and the R2 value increased by 1.24% and 1.67%, respectively. In addition, the factors affecting the failure of CFDST columns are analyzed through the parameter sensitivity analysis.

Prediction of Potato (Solanum tuberosum L.) Yield Based on Machine Learning Methods

This research delves into the application of machine learning methods for predicting the yield of potato varieties used for French fries in Poland. By integrating a comprehensive dataset comprising agronomical, climatic, soil, and satellite-based vegetation data from 36 commercial potato fields over five growing seasons (2018–2022), we developed three distinct models: non-satellite, satellite, and hybrid. The non-satellite model, relying on 85 features, excludes vegetation indices, whereas the satellite model includes these indices within its 128 features. The hybrid model, combining all available features, encompasses a total of 165 features, presenting the most-comprehensive approach. Our findings revealed that the hybrid model, particularly when enhanced with SVM outlier detection, exhibited superior performance with the lowest Mean Absolute Percentage Error (MAPE) of 5.85%, underscoring the effectiveness of integrating diverse data sources into agricultural yield prediction. In contrast, the non-satellite and satellite models displayed higher MAPE values, indicating less accuracy compared to the hybrid model. Advanced data-processing techniques such as PCA and outlier detection methods (LOF and One-Class SVM) played a pivotal role in model performance, optimising feature selection and dataset refinement. The study concluded that machine learning methods, particularly when leveraging a multifaceted approach involving a wide array of data sources and advanced processing techniques, can significantly enhance the accuracy of agricultural yield predictions. These insights pave the way for more-efficient and -informed agricultural practices, emphasising the potential of machine learning in revolutionising yield prediction and crop management.

A hybrid deep learning modeling based on lumped parameter model of coal-fired circulating fluidized beds for real-time prediction

The integration of large-scale renewable energy sources into the electrical grid has presented challenges that dynamic characteristics and mechanisms of thermal power plants are incapable of accomplishing, particularly circulating fluidized beds. Traditional mechanism models make it difficult to obtain precise parameter monitoring because of the complicated mechanism process. Accurate parameter monitoring in the circulating fluidized bed unit, assisted by the advent of artificial intelligence, significantly increases the unit's operating stability. The combination of artificial intelligence technology and mechanism model research provides a new solution to improve the accuracy of parameter monitoring. The bed temperature parameter of circulating fluidized bed units is taken as one of the most predominant characteristics of CFB operational stability. By analyzing and simplifying the bed temperature lumped mechanism model, this research identified the expression form of the feedforward neural network model and the calculation form of the convolutional neural network model and proposed a deep learning model based on the bed temperature lumped parameter model. The superior effectiveness and precision of the model are verified by the operation data of two actual units in this paper. The prediction model in this chapter achieved mean absolute error (MAE) and mean absolute percentage error (MAPE) values of 2.2579 °C, 1.5453 °C, 0.2417%, and 0.1736%, respectively. Finally, the model has also been successfully applied in the prediction task of in-situ sulfur dioxide of circulating fluidized bed units.

Impact of Covid-19 Outbreak on Turkey Electricity Generation

With this study, it was aimed to analyze the change of electricity generation in Turkey depending on COVID-19 pandemic. For this purpose, analyzes of the changes in nine different electricity generation sectors besides total electricity generation were done. In addition, these analyze were also carried out on a monthly basis in order to better understand the change in total electricity generation. Trend, naive, exponential smoothing and holt linear trend methods were used for the analysis. The results were obtained by taking into account the 2020 forecast values of the method belonging to the mean absolute percentage error (MAPE) values obtained from these methods. According to these results, it was concluded that the COVID-19 epidemic negatively affected electricity generation from total, dam, lignite, stream, imported coal and wind and did not adversely affect electricity generation from natural gas, wind, solar, geothermal and biomass in Turkey for 2020. From the results of the analysis done for the monthly change in total electricity generation in 2020, it was seen that there was a decrease in the total generation compared to the expected in the months when there were restrictions.

Improvement of Productivity in Buildings Construction

Abstract Improving productivity in construction projects has long been a major concern, and much research has been carried out to try to ameliorate construction productivity. To this end, this study aims to improve and increase the productivity rate of flat slab formwork used in residential construction projects. A survey consisting of 150 questionnaires was undertaken to identify the factors that influence on the productivity. Based on the relative Importance Index (RII), data on eleven factors deemed to affect productivity were selected. A collection of 100 data points from various sites were utilized to develop two models. Firstly, an Artificial Neural Network (ANN) model was employed, and secondly, a parametric approach was investigated. The data were divided into two sets, with 70% of the data used for training and the remaining 30% used for testing. The models’ performance was evaluated using the Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE) values. In the test phase, the artificial neural network model yielded an MSE value of 2.6610e−4 and a MAPE value of 4.9227, whereas the parametric model produced an MSE of 0.040 and a MAPE of 9.525. It was found that the artificial neural network model provided reliable prediction accuracy compared to the parametric model. However, the artificial neural network approach can be selected as a robust model in predicting and controlling the productivity rate in local construction projects by using the developed model based on the identified factors.

PERAMALAN JUMLAH KEDATANGAN WISATAWAN ASING BANDARA DI BALI DAN BANTEN MENGGUNAKAN METODE HOLT-WINTER ADITIF DAN MULTIPLIKATIF

Tourism is an important sector in the Indonesian economy. One of the benchmarks for the development of the tourism sector is the number of foreign tourist arrivals to Indonesia. Forecasting the number of foreign tourist arrivals is needed so that actors contributing to the tourism sector can optimize their service efforts. It is necessary to forecast the number of foreign tourist arrivals, especially through the arrival gate at I Gusti Ngurah Rai airport (Bali) and Soekarno-Hatta airport (Banten) as one of the main arrival gates most visited by foreign tourists. This study aims to predict the number of foreign tourist arrivals through the airport by comparing the accuracy of the Holt-Winter's additive and multiplicative method. MAPE (Mean Absolute Percentage Error) and the Durbin Watson statistical test are used and to measure the accuracy of the forecast value against the original data. Overall, the MAPE value and Durbin Watson statistical test result indicate that the additive and multiplicative approaches are good enough to be used. However, judging from the smallest MAPE value, Holt-Winter multiplicative is better used in processing data on the number of foreign tourist arrivals at both I Gusti Ngurah Rai and Soekarno-Hatta airports with MAPE values ??of 7.57% and 6.80% respectively.

Implementasi Analisis Fuzzy Time Series Dalam Melihat Hasil Komoditas Perikanan Di Provinsi Aceh

Fisheries is one of the sources of food security in Indonesia which plays an important role in the rich biodiversity of Indonesian coastal waters. This research uses the Chen Fuzzy Time Series Model forecasting method. The samples used in this study are data on the number of fishery commodity products with specific Tilapia and Catfish monthly Aceh Province as much as 24 data, starting from January 2021 to December 2022. The results of forecasting the number of tilapia in January 2023 were 1,320,815, with a MAPE (Mean Absolute Percentage Error) value of 12.93%. And the result of forecasting the number of catfish in January 2023 was 2,339,568, with a MAPE (Mean Absolute Percentage Error) value of 22.9%. This shows that forecasting using the Chen model Fuzzy Time Series method is good enough to forecast the number of tilapia and catfish in Aceh Province.

PERAMALAN JUMLAH PRODUKSI TANDAN BUAH SEGAR (TBS) KELAPA SAWIT MENGGUNAKAN METODE FUZZY TIME SERIES (STUDI KASUS: PT KALIMANTAN SAWIT KUSUMA)

Palm oil is a type of plantation crop that occupies an important position in the agricultural and plantation sectors. This is because oil palm is the largest producer of economic value per hectare in the world. The oil produced by palm oil comes from Fresh Fruit Bunches (FFB) which consist of various levels of maturity. PT Kalimantan Sawit Kusuma is experiencing instability in the amount of production, so forecasting is needed to estimate how to keep the amount of production stable. The method used is fuzzy time series (FTS), which is data forecasting that applies fuzzy sets as a basis for forecasting modeling by processing past data patterns to be applied in predicting future data. The goal is to obtain forecasting results for FFB production from April 2023 to December 2023 and obtain a Mean Absolute Percentage Error (MAPE) value. Forecasting results of Palm Oil FFB Production applying FTS Lee Order 1 in April 2023 to December 2023 are 8,412, 8,309, 8,309, 8,309, 8,309, 8,309, 8,309, 8,309 and 8,309 tons with a MAPE value of 9.07046%. While the forecasting results for Palm Oil FFB Production using FTS Lee Order 2 for April 2023 to December 2023 are 7,309, 8,559, 7,309, 8,559, 7,309, 7,309, 8,559 and 7,309 tons with a MAPE value of 4.80541%.

A new hybrid model SARIMA-ETS-SVR for seasonal influenza incidence prediction in mainland China.

Seasonal influenza is a serious public health issue in China. This study aimed to develop a new hybrid model for seasonal influenza incidence prediction and provide reference information for early warning management before outbreaks. Data on the monthly incidence of seasonal influenza between 2004 and 2018 were obtained from the China Public Health Science Data Center website. A single seasonal autoregressive integrated moving average (SARIMA) model and a single error trend and seasonality (ETS) model were built. On this basis, we constructed SARIMA, ETS, and support vector regression (SARIMA-ETS-SVR) hybrid model. The prediction performance was determined by comparing mean absolute error (MAE), mean square error (MSE), mean absolute percentage error (MAPE), and root mean square error (RMSE) indices. The optimum SARIMA model was SARIMA (0,1,0) (0,0,1)12. Error trend and seasonality (ETS) (M,A,M) was the SARIMA optimal model. For the fitting performance, the SARIMA-ETS-SVR hybrid model achieved the lowest values of MAE, MSE, and RMSE, in addition to the MAPE. In terms of predictive performance, the SARIMA-ETS-SVR hybrid model had the lowest MAE, MSE, MAPE, and RMSE values among the three models. The study demonstrated that the SARIMA-ETS-SVR hybrid model provides better generalization ability than a single SARIMA model and a single ETS model, and the predictions will provide a useful tool for preventing this infectious disease.