Holt-Winters Method Research Articles

Multiplicative Holts Winter Model for Trend Analysis and Forecasting of COVID-19 Spread in India.

The surge of the novel COVID-19 caused a tremendous effect on the health and life of the people resulting in more than 4.4 million confirmed cases in 213 countries of the world as of May 14, 2020. In India, the number of cases is constantly increasing since the first case reported on January 30, 2020, resulting in a total of 81,997 cases including 2649 deaths as of May 14, 2020. To assist the government and healthcare sector in preventing the transmission of disease, it is necessary to predict the future confirmed cases. To predict the dynamics of COVID-19 cases, in this paper, we project the forecast of COVID-19 for five most affected states of India such as Maharashtra, Tamil Nadu, Delhi, Gujarat, and Andhra Pradesh using the real-time data. Using Holt–Winters method, a forecast of the number of confirmed cases in these states has been generated. Further, the performance of the method has been determined using RMSE, MSE, MAPE, MAE and compared with other standard algorithms. The analysis shows that the proposed Holt–Winters model generates RMSE value of 76.0, 338.4, 141.5, 425.9, 1991.5 for Andhra Pradesh, Maharashtra, Gujarat, Delhi and Tamil Nadu, which results in more accurate predictions over Holt’s Linear, Auto-regression (AR), Moving Average (MA) and Autoregressive Integrated Moving Average (ARIMA) model. These estimations may further assist the government in employing strong policies and strategies for enhancing healthcare support all over India.

Designing Data Warehouse For Forecast and Data Visualization of Sales Nutrition Products

Sales data can be processed in such a way that it can become information that is used as material for analysis and consideration in making decisions. This study aims to visualize PT XYZ sales data for nutritious intake products and predict sales figures for 2018 and 2019. Data is obtained directly from PT XYZ by submitting a request for data withdrawal. Data on sales of nutritious beverage products for the last 5 years are processed using Pentaho tools with ETL method (extract, transform, load) then predicted sales figures for 2019 using R programming language with ARIMA and Holt-Winters methods after which data will be visualized using Powe BI so that the display of data presentation is more interesting and informative. To find out the compatibility in using the forecasting method, the writer will compare RSME numbers from both methods and use the method with the smallest RSME number.

FORECASTING ELECTRICITY SUPPLIED IN TURKEY USING HOLT-WINTERS’ MULTIPLICATIVE METHOD AND ARTIFICIAL NEURAL NETWORK (ANN) MODELS

Electricity is one of the most essential necessities in today’s world and has an important role for the development of societies and economics. The need for electricity is expanding continuously due to increasing population, urbanization and industrialization. Hence, the purpose of this study was to develop the best model for forecasting electricity supplied in Turkey by applying the multiplicative Holt-Winters method and multilayer feed-forward neural network model. The monthly electricity supplied in Turkey from January 2000 until December 2019 were obtained from monthly electricity statistics report presented by the International Energy Agency (EIA). The data were divided into two sets comprising in-sample data from January 2000 until December 2015 and out-sample data from January 2016 to December 2019. The multiplicative Holt-Winters was used since the electricity supplied in Turkey exhibit trend and seasonal gave the out-sample forecast of 3.6990. The best multilayer feed forward neural network (MFFNN) model with three input lag variable, one hidden node, one output node, sigmoid transfer function in hidden layer and linear transfer function in output layer gave the out-sample forecast of 2.1483. Hence it can be concluded that, the multilayer feed-forward neural network model is more accurate than multiplicative Holt-Winters method to forecast the electricity supplied in Turkey.

Teletrabalho, produção e gasto público: o que aprendemos com a covid-19?

The work investigate the behavior of variables associated with production and public spending at the State Secretariat of Finance, Espirito Santo, Brazil, between March and July 2020, considering the broad adherence to teleworking in the period - which reached 93.2% of the total workers of the institution. The phenomenon was one of the consequences of the Covid-19 pandemic, which as a choice of local public policy involved social isolation and the need to adapt to not interrupt the provision of services. The research is quantitative, using a database extracted from institutional records of the main electronic systems used by the agency, with items of expenses analyzed using the Holt-Winters method and values at constant prices. The results show that the production levels maintained an average behavior after the initial adaptation period and that the number of employees allocated in the period remained stable, with a slight downward trend. There was a reduction in spending items (water, electricity and travel) in relation to their historical values. The work contributes to research in the area by presenting an empirical study involving teleworking and public administration, in a unique context provided by the pandemic, collaborating with quantitative information on public spending and production, which corroborate with perceptions identified in previous qualitative research. Regarding the identification of evidence that can support the development of public policies, there are indications that a robust planning to rationalize the use of physical environments and structures could have caused a reduction in other items of public expenditure.

Teletrabalho, produção e gasto público: o que aprendemos com a covid-19?

The work investigate the behavior of variables associated with production and public spending at the State Secretariat of Finance, Espirito Santo, Brazil, between March and July 2020, considering the broad adherence to teleworking in the period - which reached 93.2% of the total workers of the institution. The phenomenon was one of the consequences of the Covid-19 pandemic, which as a choice of local public policy involved social isolation and the need to adapt to not interrupt the provision of services. The research is quantitative, using a database extracted from institutional records of the main electronic systems used by the agency, with items of expenses analyzed using the Holt-Winters method and values ​​at constant prices. The results show that the production levels maintained an average behavior after the initial adaptation period and that the number of employees allocated in the period remained stable, with a slight downward trend. There was a reduction in spending items (water, electricity and travel) in relation to their historical values. The work contributes to research in the area by presenting an empirical study involving teleworking and public administration, in a unique context provided by the pandemic, collaborating with quantitative information on public spending and production, which corroborate with perceptions identified in previous qualitative research. Regarding the identification of evidence that can support the development of public policies, there are indications that a robust planning to rationalize the use of physical environments and structures could have caused a reduction in other items of public expenditure.

Forecasting Dengue Outbreak Data Using ARIMA Model

Dengue fever is an internationally recognised virus that is spread by mosquitoes and can result in death. From recorded cases, Selangor has the highest rate of dengue infection among Malaysian populations. Corona disease (COVID-19), a new pandemic that has swept the globe, like Selangor, has prompted this report on the pattern of dengue cases during COVID-19 pandemic. Due to the new outbreak of COVID-19, the Movement Control Order (MCO) has been extended from time to time, and with most health resources at the state and federal levels being used to combat COVID-19, dengue control activities have been limited to a non-contact activity in outbreaks and hotspot areas. The importance of this study is to investigate the increase in dengue cases in Selangor because Selangor recorded the highest number of cases in Malaysia. Considering that there are not many studies conducted in Selangor, this study is important to predict dengue cases, and the authorities can take immediate action to overcome this problem. The aim of this research is to find the best ARIMA model for predicting the dengue cases in Selangor in the future. Several ARIMA models were used to test dengue cases data obtained in Selangor in order to achieve the objectives. The best model was calculated by comparing the Mean Square Error (MSE), Root Mean Square Error (RMSE), and Mean Absolute Percent Error (MAPE) measurement errors. Then, the predicted number of dengue cases was calculated using the best model generated. The model for which the values of criteria are the smallest is considered as the best model. Hence ARIMA (1,1,2) was found to be the best model for predicting the dengue cases data series and this model is used to predict the number of dengue cases in Selangor with the smallest Mean Square Error (MSE) value of 12837.4327 and Root Mean Square Error (RMSE) value of 113.3024. The forecasted values showed a decreasing number of dengue cases. This study was carried out using R- studio software and excel. Further research can be conducted using another time series method, for example Holt-winters method.

Modeling and forecasting of principal minerals production

Although coal reserves are abundant in Pakistan, still share of gas and oil is about 65% in the energy mix. Pakistan, despite being a mineral-enriched country, is facing an alarming situation as its power generation is based on foreign exchange. The mineral sector of Pakistan is dominated by four principal minerals which are gas, oil, gypsum, and coal, while gypsum being a source of reclamation and poverty reduction. There is a strong need to analyze and forecast the production of these four principal minerals to cope up with emerging challenges. Data contains 114 observations from the period of July 2005 to December 2014, measured in terms of metric tonnes (MT). In parametric models, Box-Jenkins (B-J) methodology, a regression model with auto-regressive errors (ARAR), and Holt-Winter (HW) method are used to model. In nonparametric models, univariate singular spectrum analysis (SSA) and multivariate SSA (MSSA) modeling approach are applied. Data is split into train and test data in order to specify a suitable model for forecasting. Root Mean Square Error (RMSE), Mean Absolute Percentage Error, and Theil’s U statistic are utilized as the measure of accuracy. For gas and coal, HW model is a suitable model to forecast. For gypsum and oil, Auto-regressive Integrated Moving Average (Box Jenkins ARIMA) and MSSA provide more accurate predictions, respectively. The forecasts for gas and gypsum as compared to 2014 are expected to be approximately 11 % and 45 %, respectively, more in 2020. In 2020, the forecasts of oil are expected to be eight times more than in 2014. The production of coal in 2020 is expected to decrease 12 % than in 2014. There is a strong need to optimize the production of coal by providing incentives for exploration and mining. The stakeholders should make serious efforts to bring the production of coal at an optimum level such as by providing modern equipment and high incentives to promote coal mining and exploration.

ANALYSIS OF LOAD FORECASTING ACCURACY BASED ON HO CHI MINH CITY DATA

Short term load forecasting is one of the fundamental parts of the electric system. Among exponential smoothing methods, the Holt-Winters method is widely used to forecast the short-term load since it is easy and simple to use, and it has high ability to adapt to the forecast of different time horizons. This paper presents a new approach by combining Holt-Winters and Walk-Forward Validation methodology to forecast the maximum power demand for Ho Chi Minh City, Vietnam. The data is divided into the training and test sets in many cases. The forecast accuracy of the mean absolute error (MAE) and the mean absolute percentage error (MAPE) are used to analyze the characteristic of forecast for each day of the week.

Improving the precision of modeling the incidence of hemorrhagic fever with renal syndrome in mainland China with an ensemble machine learning approach.

Hemorrhagic fever with renal syndrome (HFRS), one of the main public health concerns in mainland China, is a group of clinically similar diseases caused by hantaviruses. Statistical approaches have always been leveraged to forecast the future incidence rates of certain infectious diseases to effectively control their prevalence and outbreak potential. Compared to the use of one base model, model stacking can often produce better forecasting results. In this study, we fitted the monthly reported cases of HFRS in mainland China with a model stacking approach and compared its forecasting performance with those of five base models. We fitted the monthly reported cases of HFRS ranging from January 2004 to June 2019 in mainland China with an autoregressive integrated moving average (ARIMA) model; the Holt-Winter (HW) method, seasonal decomposition of the time series by LOESS (STL); a neural network autoregressive (NNAR) model; and an exponential smoothing state space model with a Box-Cox transformation; ARMA errors; and trend and seasonal components (TBATS), and we combined the forecasting results with the inverse rank approach. The forecasting performance was estimated based on several accuracy criteria for model prediction, including the mean absolute percentage error (MAPE), root-mean-squared error (RMSE) and mean absolute error (MAE). There was a slight downward trend and obvious seasonal periodicity inherent in the time series data for HFRS in mainland China. The model stacking method was selected as the best approach with the best performance in terms of both fitting (RMSE 128.19, MAE 85.63, MAPE 8.18) and prediction (RMSE 151.86, MAE 118.28, MAPE 13.16). The results showed that model stacking by using the optimal mean forecasting weight of the five abovementioned models achieved the best performance in terms of predicting HFRS one year into the future. This study has corroborated the conclusion that model stacking is an easy way to enhance prediction accuracy when modeling HFRS.

Cloud-Fog Assisted Energy Efficient Architectural Paradigm for Disaster Evacuation

The disasters often lead to panic among people resulting into life threatening crowd movements. An efficient and systematic evacuation can mitigate the impact of devastation on human life. The evacuation, on the wake of extreme events, requires delay-sensitive computation and energy conserving mechanisms for power constrained data collection devices. In this paper, an IoT based Cloud-Fog assisted layered evacuation system is proposed for mass evacuation during emergency situation. The system utilizes fog computing paradigm to evaluate the panic health state in real-time using fuzzy K-nearest neighbour (FK-NN) method. The energy conserving mechanisms at fog layer performs data selection and data reduction to reduce the amount of data transmission to upper layer. Spatio-temporal Data Selection layer chooses the data for further transmission on the basis of its spatio-temporal analysis. Data reduction component reduces the dimensions of data using Principal Component Analysis (PCA). The cloud layer examines the future temporal health state of evacuees by employing Holt-Winters method. Geographic Analysis layer at cloud determines the priority evacuation status of person on the basis of current and future temporal panic state. Moreover, it specifies the priority evacuation status of region using the count of currently panicked and predicted health status. Experimentation proves the efficiency of proposed system at various levels.

Power load forecasting research based on neural network and Holt-winters method

Since the electric load shows a very obvious periodicity in time series, increasing the periodic factor in the power load forecasting is a research direction of power load forecasting. It is by adding trend and seasonality (that is, periodicity) to smoothing values that the Holt-winters algorithm improves the accuracy of predictions. In this paper, Holt-winters algorithm and neural network algorithm are used to build power load prediction models respectively, and data of city A in Shandong province is used for testing. Experimental results show that the prediction results of the two algorithms are similar, but the Holt-winters algorithm is slightly more accurate.

Univariate Time Series Prediction of Wind speed with a case study of Yanbu, Saudi Arabia

Wind energy is a promising alternativefor renewable source of energy pursued world-wide to reduce carbon emissions for a green future. The prediction of wind speed is a challenging subject and plays an instrumental role in development of wind power systems (particularly grid connected renewable energy systems where predicting wind speed facilitates manipulation of the load on the grid). Modern machine learning techniques including neural networks have been widely utilized for this purpose. Literature indicates availability of several models for estimation of the wind speed one hour ahead and the hourly wind speed data profile one day ahead. This paper considers the prediction of wind energy as a univariate time series (UVT) prediction problem and employs major prediction algorithms including the K-Nearest Neighbors (kNN), Random Forest (RF), Support Vector Regression (SVR), Holt-Winter and ARIMA method. Forecasting a univariate time series depends only on past wind speed data values, rather than use of external data attributes like wind direction or weather forecast for prediction algorithm. In the present study (as a case-study), 13 years of hourly average wind speed data (of the period 1970-1982) of Yanbu, Saudi Arabia has been utilized to evaluate the performance of selected algorithms. Yanbu is an industrial city that plays a major role in the economy of Saudi Arabia. The findings showed that SVR, RF and ARIMA methods exhibit a better forecastingperformance in relation to four evaluation parameters of Mean Absolute Percentage Error(MAPE),Symmetric Mean Absolute Percentage Error (sMAPE),Mean Absolute Error (MAE) and Mean Absolute Scaled Error (MASE).

Scalable Wireless Wearing Monitoring System for Harsh Industrial Environment

A scalable system for wireless wear monitoring in a distributed harsh industrial environment, in which installation of the electronic equipment near the distributed wearing measuring locations is unfeasible, has been successfully designed, developed and deployed. The system is composed of a novel <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LC</i> -type passive wireless wear sensor system, a sensor readout coil, an electronic equipment, a readout unit for a scalable multisensor system and a control and monitoring system. The capacitor structure, employed as a sensing device in the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LC</i> sensor manufactured on low cost PCB, provides a linear response to the wear level. This allowed us to define a measuring and calibration method to suppress the cable effects from each distributed readout coil to the electronic equipment readout unit. It was possible to achieve a multiple simultaneous measurements in different sensors distributed along an industrial process without significant cabling influence. Concerning the sensor data analysis, a double Holt-Winters method with trend was used to reduce noise effects and an “experience based learning” tool was implemented in order to solve any misbehavior or time lag. Laboratory measurements and on-site operation results in an article mill indicate 1 mm precision and about 0, 5 MHz/mm sensitivity of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LC</i> -type passive wireless wear sensor system.

Flow-based profile generation and network traffic detection for DNS anomalies using optimised entropy-based features selection and modified Holt Winter's method

Flow-based profile generation and network traffic detection for DNS anomalies using optimised entropy-based features selection and modified Holt Winter's method

СИСТЕМА МОНИТОРИНГА ПОКАЗАТЕЛЕЙ ДОСТУПНОСТИ СЕТЕЙ МОБИЛЬНОЙ СВЯЗИ СТАНДАРТА LTE-A

In this paper we consider two accessibility indicators, namely E-RAB (E-UTRAN Radio Access Bearer) and E-RRC (Evolved Radio Resource Control) failure rates, of the LTE-A communication network belonging to one of the regional operators in Russian Federation. The aim of this study is to find the proper algorithms for accessibility indicators prediction, and performance estimation of these algorithms. During the study, we provide temporal dynamics of the indicators and possible failure reasons, behind these indicators. Then the percentage of the time series values is shown, which are corresponding to the abnormal situations (incidents). After that, the stationarity of the inspected time series using augmented Dickey-Fuller (ADFuller), and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) methods is analyzed. Next, the ETS decomposition is performed. In order to predict the future values of the indicators, we utilize SARIMA model, triple Exponential Smoothing (Holt-Winters method), Facebook Prophet, Prony decomposition based model and XGBoost algorithm. Performance estimation is obtained in two ways: by test-sequence- and cross-validation-based Median Absolute Error (MAE). Also, the architecture for the monitoring system, that collects, analyzes and visualizes the required metrics within the infrastructure of the considered operator, is proposed in this paper. Herein, we analyze the possibilities of the open-source solution deployment on each stage of the monitoring process from data mining and preparation up to predictive model learning.

FORECASTING OF SOME KEY INDICATORS OF THE RFI AND RFP PROCESSES OF THE BULGARIAN MOBILE TELECOMMUNICATION OPERATORS

The present paper regards the opportunities of forecasting of some key indicators in the “Request for Information” (RFI) and “Request for Proposal” (RFP) processes in the supply chain at the Bulgarian mobile telecommunication operators. The presented hereby forecasting is based on the use of the Holt-Winters method for exponential smoothing in the presence of additive and multiplicative seasonality and is made or indicators: “number of contracts”, “number of contracts with savings” and “number of the issued purchase orders”. The lowest “Stationary R square”, “R square” and MAPE (Mean Absolute Percentage of Error) values are used as measurement of accuracy and for selection of the best fit models that are applied. It is also important to point out that the measurement is being done for the so-called “bottle necks” or “narrow places” in the RFI and RFP processes. The purpose of this bottle-neck forecasting is to provide timely point for “Go/Not Go” decisions point for these very same process and thus to result in an improved risk management in the form of risk aversion and risk minimization.

An Integrated Model for Forecasting Indian Automobile

The automobile industry is one of India’s main economic sectors. In recent decades India has attracted many global players in the automobile industry. The industry has significantly benefited from an increase in the paying capacity of the consumers. This has contributed to increased competition in the market. Given that the automobile industry is a very complex process, a tool to predict the future of automotive demand from the modeling point of view is needed because of its high level of complexity and uncertainty. This study aims to introduce a novel integrated model with a combination of Adaptive Multiplicative Triple Exponential Smoothing Holt-Winters (AHW) method and Backpropagation Neural Networks (BPNNs) to improve the likelihood of predicting automobile sales accurately. This study is subject to continue validating a model in real-world automobile selling data against existing methods. This model also incorporates the linear and non-linear characteristics of AHW and BPNN, respectively to form a synergistic model. The proposed model has the higher capability to provide reasonable accuracy in forecasting future sales in terms of average prediction accuracy of 0.974637 than the existing methods namely BPNN 0.9483 and ANN 0.9275. For training and testing purposes, validation is done using the Indian automobile sales data. Finally, the regression fit shows that during the test stage in the car sales data for the period 2016-2017 and 2017-2018, the proposed integrated model is better than the conventional method.

Time series prediction of under-five mortality rates for Nigeria: comparative analysis of artificial neural networks, Holt-Winters exponential smoothing and autoregressive integrated moving average models

BackgroundAccurate forecasting model for under-five mortality rate (U5MR) is essential for policy actions and planning. While studies have used traditional time series modeling techniques (e.g., autoregressive integrated moving average (ARIMA) and Holt-Winters smoothing exponential methods), their appropriateness to predict noisy and non-linear data (such as childhood mortality) has been debated. The objective of this study was to model long-term U5MR with group method of data handling (GMDH)-type artificial neural network (ANN), and compare the forecasts with the commonly used conventional statistical methods—ARIMA regression and Holt-Winters exponential smoothing models.MethodsThe historical dataset of annual U5MR in Nigeria from 1964 to 2017 was obtained from the official website of World Bank. The optimal models for each forecasting methods were used for forecasting mortality rates to 2030 (ending of Sustainable Development Goal era). The predictive performances of the three methods were evaluated, based on root mean squared errors (RMSE), root mean absolute error (RMAE) and modified Nash-Sutcliffe efficiency (NSE) coefficient. Statistically significant differences in loss function between forecasts of GMDH-type ANN model compared to each of the ARIMA and Holt-Winters models were assessed with Diebold-Mariano (DM) test and Deming regression.ResultsThe modified NSE coefficient was slightly lower for Holt-Winters methods (96.7%), compared to GMDH-type ANN (99.8%) and ARIMA (99.6%). The RMSE of GMDH-type ANN (0.09) was lower than ARIMA (0.23) and Holt-Winters (2.87). Similarly, RMAE was lowest for GMDH-type ANN (0.25), compared with ARIMA (0.41) and Holt-Winters (1.20). From the DM test, the mean absolute error (MAE) was significantly lower for GMDH-type ANN, compared with ARIMA (difference = 0.11, p-value = 0.0003), and Holt-Winters model (difference = 0.62, p-value< 0.001). Based on the intercepts from Deming regression, the predictions from GMDH-type ANN were more accurate (β0 = 0.004 ± standard error: 0.06; 95% confidence interval: − 0.113 to 0.122).ConclusionsGMDH-type neural network performed better in predicting and forecasting of under-five mortality rates for Nigeria, compared to the ARIMA and Holt-Winters models. Therefore, GMDH-type ANN might be more suitable for data with non-linear or unknown distribution, such as childhood mortality. GMDH-type ANN increases forecasting accuracy of childhood mortalities in order to inform policy actions in Nigeria.

Forecasting the server status using the triple exponential smoothing model

This article examines the issue of forecasting the time series of server systems and suggests applying the triple exponential smoothing model to solve this problem. It presents a mathematical formulation of the problem and describes the specifics of forecasting the time series of server systems. After this the article gives a comparative analysis of the autoregressive, neural network and exponential smoothing models in terms of their application to this problem. It argues that the triple exponential smoothing model (Holt-Winters method) offers a number of advantages when modelling the time series of server systems. It then provides experimental research to evaluate the accuracy of the Holt-Winters method with respect to the indicated time series. The research shows that the triple exponential smoothing model exhibits high results and can be applied to the solution of practical problems.

Application of machine learning time series analysis for prediction COVID-19 pandemic

PurposeCoronavirus disease is an irresistible infection caused by the respiratory disease coronavirus 2 (SARS-CoV-2). It was first found in Wuhan, China, in December 2019, and has since spread universally, causing a constant pandemic. On June 3, 2020, 6.37 million cases were found in 188 countries and regions. During pandemic prevention, this can minimize the impact of the disease on individuals and groups. A study was carried out on coronavirus to observe the number of cases, deaths, and recovery cases worldwide within a specific time period of 5 months. Based on this data, this research paper will predict the future spread of this infectious disease in human society.MethodsIn our study, the dataset was taken from WHO “Data WHO Coronavirus Covid-19 cases and deaths-WHO-COVID-19-global-data”. This dataset contains information about the observation date, provenance/state, country/region, and latest updates. In this article, we implemented several forecasting techniques: naive method, simple average, moving average, single exponential smoothing, Holt linear trend method, Holt-Winters method and ARIMA, for comparison, and how these methods improve the Root mean square error score.ResultsThe naive method is best suited as described over all other methods. In the ARIMA model, utilizing grid search, we recognized a lot of boundaries that delivered the best-fit model for our time series data. By continuing the model, future predictions of death cases indicate that the number of deaths will increased by more than 600,000 by January 2021.ConclusionThis survey will support the government and experts in making arrangements for what is about to happen. Based on the findings of instantaneous model, these models can be adjusted to guide long time.