Search

The nonparametric regression that has been widely used in recent years is a spline. Spline has a very good statistical interpretation and visual interpretation. Several studies that have been developed by researchers are uniresponse spline truncated models. This model is very good in terms of modeling smooth data but it has not been able to solve modeling problems that contain more than one response. Therefore, the researchers began to develop a bi-response spline truncated. Three things need to be considered in the use of the spline, namely determining the number of the order, the number of the knots, and the location of the knots. Cross-Validation (CV) and Generalized Cross-Validation (GCV) has developed to select the optimal knot points. However, until now there has been no research that concludes the best method between CV and GCV. This study will compare the CV and GCV methods in selecting the optimal knot point on the poverty percentage and poverty severity index in East Java using the spline truncated bi-response model. The criteria for the selected model are based on the coefficient of determination (R2). The results of the analysis show that by using the CV method, there are 3 optimal knots with R2 value of 94.71%. Meanwhile, using the GCV method, there are 3 optimal knots with R2 of 95.76%. From these results, it can be said that GCV is the best method for choosing the optimal knot point compared to the CV method. [ABSTRACT FROM AUTHOR]

This study aims to determine the best equation method and model for prediction unemployment in North Sumatra in 2022. This study uses dataset from the Central Statistics Agency (BPS) of North Sumatra where dataset is taken for 22 years from 2000 until 2021. The results of this study are the best parameters of each method were for the Double Moving Average method with an average time parameter of 3rd periods and the Double Exponential Smoothing method with alpha value = 0.5. The 3rd-period Double Moving Average accuracy test shows a MAPE value of 13.97% while Double Exponential Smoothing with alpha value=0.5 indicates a MAPE value of 16.45%. The study shows that the best forecast model for North Sumatra unemployment in 2022 is to use the 3rd period using the DMA method to forecast 2022 unemployment up to 519305 people. [ABSTRACT FROM AUTHOR]

Exponential smoothing models are used to smooth time series data and treat the impact of Trend or Seasonal in the time series. There are several models including of smoothing: Simple Exponential Smoothing, Double Smoothing and Holt-Winters smoothing. Robust statistical methods in time series models or in linear models (Regression, Experiment Design) they are a suitable to treat the problem of outlier values in data in order to obtain more reliable forecast results. In this paper, we applied the Robust Exponential smoothing method Compared with some anther exponential as :(Double smoothing, Holt-Winters Additive/Multiplicative smoothing, EST method) for estimation model parameters and forecasting are presented. and application of all these methods in Data of corona Pandemic (COVID-19) from Iraq in period (Mar. 2020 – Mar 2022), the robust exponential method is presented that applies the standard technique to preclean data from the outlier values in time series.We depended on R-packages functions as [ses(),EST, robiest(), hw(),...],it use to obtain smoothing parameters are robust. In this research we found Robust method have a good forecast performance for time series with an outlier values. [ABSTRACT FROM AUTHOR]

Rainfall is still a prominent meteorological phenomenon that has piqued the interest of many scientists who have sought to forecast and quantify its future proportions. Malaysia has had multiple devastating floods in recent years. Rainfall forecasting is critical for watershed management applications, particularly flood warning systems. Peninsular Malaysia saw its most recent flood in December 2021, affecting seven states. In this work, we will use the Box-Jenkin model, exponential smoothing, and a hybrid model that combines the Box-Jenkin and exponential smoothing approaches to estimate rainfall in Senai, Johor. Senai, Johor, was chosen due to its favourable location midway between east and west Malaysia. As a result, no strong monsoons, such as those in the northeast or southwest, occurred. Mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE), and a graphical representation will be used to compare the performance of the models. The results show that the Holt-model Winter's performs the best in predicting rainfall in Senai, Johor. The hybrid model, which combines two conventional models, did not enhance prediction, proving that it is not always required to create a hybrid model. [ABSTRACT FROM AUTHOR]

One of the measurements for evaluating time series forecasting performances is the mean square error (MSE). This paper proposes an algorithm to find the smallest MSE. To capture the four components of the time series data (namely, seasonal variation, trend variation, cyclical variation, and random variation), the exponential smoothing method was used. This method uses three factors for smoothing where the data are the smoothing factor, 0 < α < 1, the trend smoothing factor, 0 < γ < 1, and the seasonal change smoothing factor, 0 < β < 1. First, all possible combination values of smoothing factors will be generated to one decimal digit. After that, the smallest MSE of those combinations will be determined by using an algorithm and observing their convergence pattern. [ABSTRACT FROM AUTHOR]

Spline functions and digital filtering are two numerical methods normally used by the researchers in the field of biomechanics to perform data smoothing and time differentiation. The objective of this study is to perform data smoothing on the raw kinematic data of a subject performing standing-bending-lifting activity using spline function, and to calculate time differentiation from the kinematic data. Raw data on standing-bending-lifting movement was recorded in Biomechanics Lab, UniMAP. Next, quintic spline function was used to smooth the data and calculate the angular velocity and angular acceleration at L5, L3, L1, MAI and T2. When calculated for one trial, the error estimation between smooth and raw data at L5, L3, L1, MAI, and T2 are 0.008°, 0.003°, 0.003°, 0.004°, and 0.008°, respectively. When calculating for five trials, the average error estimation between smooth and raw data at L5, L3, L1, MAI, and T2 are 0.007°, 0.009°, 0.011°, 0.012°, and 0.012°, respectively. The result shows that the quintic spline is able to produce satisfactory output in data smoothing. [ABSTRACT FROM AUTHOR]

In this work, we consider a model for system of strongly interacting massless particles that interact with each other via gravitational interaction, known as the massless Einstein-Skyrme system, where the particles is considered as topological solitonic wave. In contrast with its steady state (static) limit, The dynamical state of this system is not yet well-understood. Here, we study the local existence of unique regular solutions of dynamical massless Einstein-Skyrme system as one of the important property in the study of dynamical system's well-posedness. The result is that for every smooth initial data, h0(r), that behaves like h0(r) = O(r−2) and ∂ h 0 (r) ∂ r = O (r − 3) as r→ ∞, there exist an interval of time where unique regular solution exist for that specific initial data. [ABSTRACT FROM AUTHOR]

X Islands is a natural marine tourism destination with plenty of favorite tourist spots. To go to X Islands visitors must board a passenger ship from the port. The design of the monitoring dashboard and prediction of the number of ship passengers is conducted to help the port manager monitor the number of ship passengers going to X Islands. The method used to predict ship passengers is the data mining process using the Time Series Analysis method, namely Triple Exponential Smoothing. The results obtained from this research are the formation of a data mart to store the analysis result data which will be displayed on the dashboard to help port managers read the information available on the data mart. [ABSTRACT FROM AUTHOR]

When investing in housing, the most important thing for investors to consider is the estimated price increase. So far, to predict the price increase, potential investors have made speculations without reasonable logical calculations. Given the importance of this indicator, the purpose of this design is to assist investors in choosing the type of houses with the potential to sell at a high price. This system will be designed using Double Exponential Smoothing method to assist in the process of forecasting the value of a housing investment within a predetermined time. The data used is the document on the sale price of housing in the previous period. The result of study is an information system in the form of a dashboard display that shows data on house price increase from the previous period. Investors can draw conclusions in choosing the type of house with high price increase through charts or diagrams that are presented on the dashboard. [ABSTRACT FROM AUTHOR]

Forecasting is a method for predicting future events based on time series data. Most forecasting is done for daily, weekly, monthly, even yearly. Forecasting the Air Pollutant Standard Index (ISPU) in DKI Jakarta can be used for future decision making and preventive measures. The method used is the triple exponential smoothing method from Holt-Winter with alpha = 0.4, beta = 0.01, and gamma = 0.05. Forecasting results are tested by measuring the error rate of the MAPE value and producing a MAPE of 30%, so that the Holt-Winter method has an accuracy of 70%. Forecasting results are visualized on a dashboard for use by local authorities. [ABSTRACT FROM AUTHOR]

The aim of current research is to propose an approach for evaluation the cut surface and structure of white brined cheese using methods of computer vision. The segmentation is preferred in current research in order to analyse basic structural elements on the cheese cut surface such as technical holes, layers, porcelain structures on the broken surface and etc. Since the broken surface has complex 3D structure it is not enough to use small number of colours in segmented image and because of this the simplest type of segmentation-thresholding could not extract all structural elements. It is chosen the segmentation algorithm "Statistical region merging" (SRM). Due to complexity of images of cheese samples the application of SRM algorithm leads to producing of images with too many colours. Smoothing of images by application of filters could be a preliminary step before segmentation in order to influence on the regions' growing. Two popular filters (mean and median) for images smoothing are used in current research for pre-processing the images of cheese samples before segmentation with SRM algorithm. The filters are applied with variety of kernel sizes and all filtered images are segmented. The ratio of number of colours in original image to number of colours in segmented image is calculated in percent (Kseg). This ratio corresponds with experts' estimation for quality of "Cut surface and structure" which is confirmed by correlation analysis. The results show that smoothing with mean or median filter prepares images for segmentation with SRM algorithm so that important structural elements are emphasized. An equation for calculation the estimate for cheese quality of "Cut surface and structure" using Kseg is proposed. The average absolute error for calculated estimation using proposed equation versus experts' estimation is about 17% but minimum absolute error is 7%. A benefit of proposed approach is that it not use conversion in other colour systems (such as HSI) and thus it is faster than other known techniques still it preserves strong correlation with experts' estimation (Person's coefficient is about 0,78). [ABSTRACT FROM AUTHOR]

Rainfall plays an important role for cultivation and drinking purpose. If rains are high human beings face problems like floods and rains are low human beings again face another problems like droughts. In this paper, annual rainfall data from 1988 to 2017 is used for estimation of outliers using control limits and specification limits. After eliminating outliers for modified data of Rayalaseema, Coastal Andhra Pradesh and Telangana regions, we are fitting general linear model, Holt's linear exponential smoothing model and Brown's linear exponential smoothing model. Best among these three models is estimated using R2 criteria. [ABSTRACT FROM AUTHOR]

In a Wireless Body Area Network (WBAN), the Proactive-Transmission power control algorithm is implemented by closed-loop control with posture and motion detection. It takes only the motion state of the user. The proposed Heterogenous Movement Detection based Transmission Power Control algorithm (HMD-TPC) uses hypothesis testing. This hypothesis testing predicts the activity of the user and allocates power based on exponential smoothing and neighbor node activity. The accelerometer value and Received Signal Strength (RSSI) of the gait cycle are taken as input. This work also minimizes Intra-WBAN interference up to 85%. using the emergency factor, neighbor node's activity, residual energy, and distance. Therefore, HMD-TPC provides a high packet delivery ratio and increases the percentage of interference mitigation. The transmission power level is reduced up to -15dBm compared to the existing works. [ABSTRACT FROM AUTHOR]

Forecasting is a science used to predict future events, forecasting activities are business functions used to predict sales or use of a product so it will be more efficient. Therefore, in a business, a forecasting system is needed to help predict the number of sales of an item, so that prediction can be used to estimate the supply of goods to be purchased or a profit. One of the forecasting methods that can be used in forecasting sales of goods is Holt's Double Exponential Smoothing. This forecasting technique is used on data that has a trend but is not affected by seasons. In this method, 2 weight smoothing factors are used, namely α for level smoothing weights and β for weight smoothing trends From the results of this study using 28 months of data, from November 2018 to February 2021 obtained from Jaya Mandiri which is engaged in selling plywood. The prediction error value obtained using the Mean Absolute Percentage method during the train process was 4.651% while during the validation process an error was 4.989%. For alpha parameter with a value of 0.8 and beta of 0.2, the alpha and beta parameters used in the calculation have been searched automatically and the optimal alpha and beta values. The conclusion of this study are Holt's Double Exponential Smoothing can be applied to these cases because the MAPE produced is below 10% both during processing and validation. This application is also successful in determining the optimal alpha and beta parameters. [ABSTRACT FROM AUTHOR]

The purpose of this research is to compare forecasting methods to find out the best forecasting method in predicting the sales value of Concentrated Feed based on the error smallest. The method used is the method of Moving Average, Single Exponential Smoothing and Trend linear. Based on the selection of the Mean Absolute Percentage Error or can be abbreviated as MAPE, the Moving Average uses a three period and five period. In the Moving Average with a three-period moving average, the MAPE value is 2.34%, and in the Moving Average with a five-period moving average, the MAPE value is 2.36%. Then the Exponential smoothing used values, α 0.5 and 0.9. In the Exponential smoothing with used values, α=0.5 the MAPE value is 2.07%, and in the Exponential smoothing with used values, α=0.9, the MAPE value is 2.26%. In the linear trend method, the MAPE value is 1.75%. After comparing the five forecasting methods used, the linear trend method was chosen due to its lower error rate compared to both the Moving Average method and Single Exponential Smoothing. [ABSTRACT FROM AUTHOR]

We consider the extremal problem of interpolation of convex scattered data in ℝ3 by smooth edge convex curve networks with minimal Lp-norm of the second derivative for 1 < p ≤ ∞. The problem for p = 2 was set and solved by Andersson et al. (1995). Vlachkova (2019) extended the results in (Andersson et al., 1995) and solved the problem for 1 < p < ∞. The minimum edge convex Lp-norm network for 1 < p < ∞ is obtained from the solution to a system of nonlinear equations with coeffificients determined by the d ata. The solution in the case 1 < p < ∞ is unique forstrictly convex data. The approach used in (Vlachkova, 2019) can not be applied to the corressponding extremal problem for p = ∞. In this case the solution is not unique. Here we establish the existence of a solution to the extremal interpolation problem for p = ∞. This solution is a quadratic spline function with at most one knot on each edge of the underlying triangulation. We also propose suffificient conditions for solving the problem for p = ∞. [ABSTRACT FROM AUTHOR]

Government debt is a flow variable that equals the difference between government receipts and expenditures in a single year, whereas government expenditure is related with multiplier effects that could potentially contribute to output increases for economic systems. When the government expenditure exceeds its revenue, a budget deficit will occur. If Malaysian government finances a budget deficit through borrowing, the Malaysian economy will be unable to generate significant and long-term economic growth. Higher deficits and debt will increase employment rate, consumption, expenditure, and eventually economic growth if they are correctly measured. Hence, it is important to choose the best model that can forecast the future value of debt and expenditure. Therefore, the main goal of this study was to identify the pattern of data and to evaluate the best model for estimating the future value of debt guaranteed and spending of the Federal Government of Malaysia in 2019. The debt and expenditure data were obtained from Bank Negara Malaysia for the year 2013 to 2018. In overall, the guaranteed debt had been increasing over time, while the government expenditure fluctuated over time and had seasonal behavior. The Granger causality test showed that government debt caused the future performance of expenditure. The results also suggested that double exponential smoothing was the most suitable model in forecasting debt guaranteed of the federal government. Meanwhile, the best model suggested to forecast federal government's expenditure by was SARIMA model due to the presence of the seasonality effect. [ABSTRACT FROM AUTHOR]

Liver disease is a disease that affects many humans. Machine learning that is developing at this time, is also widely implemented in the medical field. In this study, machine learning was used to estimate liver disease. To improve the performance of the estimation results, data pre-processing is carried out in the form of data transformation with exponential smoothing (ES). From the experiments carried out, it turns out that the results of the machine learning performance increase. And the lowest RMSE results were obtained from the optimization of the knn method and the best average RMSE was obtained from the neural network (NN) method. [ABSTRACT FROM AUTHOR]

Time-series imagery data provides an opportunity to explore the seasonality parameters during the growing seasons of crops. Unfortunately, the quality of time-series data often gets disturbed by outliners and noisy pixels. One of the solutions to this problem is using the TIMESAT program to smooth the time-series data and generate seasonality parameters. Several studies used TIMESAT software to enhance their research data quality. As one of the popular methods in smoothing time-series data and generating seasonality parameters, this paper aims to present the description of the TIMESAT program, provide explanations of the TIMESAT working system, and elaborate the setting of TIMESAT parameters. TIMESAT program consists of several processing logics which provide sequence steps in data processing, making it easy to operate. However, running this program could be challenging in terms of determining the appropriate fitting algorithm and appropriate parameter settings. This paper explores the provided guideline to set TIMESAT parameters and uses one of the previous studies on TIMESAT to explain the process of determining each TIMESAT parameter. To sum up, setting TIMESAT parameters involved trial and error in the range of values provided by the guideline. Once the fitted curve looks appropriate in representing the original time-series data, we can conclude that we have chosen the right parameters. [ABSTRACT FROM AUTHOR]

The percentage of electrical energy needs is the largest demand compared to other energy needs such as natural gas, fuel, and coal. This is due to various factors, including population growth, economic growth, industrial development, as well as the rapid development of electricity-based technology in almost every sector, especially in the household, industrial and commercial sectors. This condition has the potential to trigger an electricity crisis, but can be minimized if the required electricity consumption is known. One way to determine electricity demand is to build a predictive model that is accurate and flexible and able to accommodate the complexity of seasonal patterns, both seasonal in months and years as reflected in electricity consumption data. Therefore, in this study, the TBATS model was used to accommodate this. TBATS models will be used which is the development of the exponential smoothing model that can accommodate the occurrence of multiple seasonal patterns, both nested and non-nested, non-integer seasonal periods, and handle the possibility of non-linearity cases because it has a flexible seasonality The results of this study, the TBATS model built has a value of SMAPE by 8.13% has been able to capture fluctuating patterns in seasonal periods. [ABSTRACT FROM AUTHOR]

The increasing population growth, of course, has a correlation with the increasing consumption of the amount of clean water needed by the community. Water production management companies need to increase the amount of water production in order to avoid the problem of shortages in meeting water needs which have a detrimental impact on consumers. The emergence of the problem factor in meeting the need for clean water makes the need for a solution to make a decision-making model in the production of the amount of clean water in the future so that adjustments can be made to the amount of water production with the demand for water needs. The data used in this study uses actual data within the previous five years, namely from 2015 to 2019 using the Double Exponential Smoothing (DES) and Triple Exponential Smoothing (TES) calculations using nine alpha values. Forecasting accuracy method used the Mean Absolute Percentage Error (MAPE) method to calculate the percentage error for each alpha value. Based on the results of data processing, comparison testing and verification of results, it shows that the DES method is more superior as a recommended method that can be used to perform calculations as decision support by companies in making decisions to predict the amount of clean water in the future. This is concluded based on the smallest MAPE value in the DES method at alpha 0.4 and the verification results from the tracking signal value which shows the value does not exceed the UCL (Upper Control Limits) and LCL (Lower Control Limits) values in the DES method compared to the TES method. [ABSTRACT FROM AUTHOR]

Cancer incidence is defined as the number of new cases of cancer in a specific population for a specific period. In Malaysia, the incidence rates for cancer keep increasing each year especially the top three cancers namely female breast, colorectal and lung cancer. The increasing trend has impacted the society economically and psychosocially. The cost of treating cancer and the efforts to psychologically manage the emotions of the patients are crucial as the patient's financial, social and physiological situations may all change accordingly. Having an informed estimations about future life with cancer is helpful in preparing for better lives. Through earlier indication of cancer predictions, effective resource planning and attentive mental wellbeing of future healthcare frameworks can be well anticipated. Therefore, having the right model to accurately forecast the incidence rates is deemed essential. This study aims to determine the best model between Box-Jenkins ARIMA and Exponential Smoothing in forecasting the incidence rates. The model with the least Mean Absolute Percentage Errors (MAPE) value, is determined as the best model and used to forecast the top three cancer incidences for the year 2017 to 2021. This forecast period is in accordance to the periodical report of Malaysia National Cancer Registry (MNCR) published by the Ministry of Health (MOH), Malaysia which is yet to be published. Hence, these forecast figures will backtest the upcoming MNCR 2017-2021 report. Results show that the Exponential Smoothing model predominantly outperformed the ARIMA model. Only for Female Colorectal Cancer, ARIMA model performs better than Exponential Smoothing. [ABSTRACT FROM AUTHOR]

Forecast future crude oil stock price trends are essential for the purpose to assist the government, economists, and investors. However, the non-trading days will affect the accuracy of forecasting. Thus, multiple imputations were applied in this study toward the non-trading days in the stock price. To verify the accuracy of multiple imputations, this study attempt to compare the forecast accuracy of both artificial neural network (ANN) and simple exponential smoothing (SES) in forecasting the future crude oil stock price with imputation and without imputation process. Based on the results, ANN with imputation consist of fifteen lags as input neurons and ANN without imputation consist of five lags input neurons. In overall, ANN outperformed SES and all forecast models with imputation outperformed the models that without imputation. Therefore, the best approach is ANN with imputation after Min-Max normalization. Thus, it can be concluded that imputation can improve the forecast accuracy and the non-trading days should not be neglected. [ABSTRACT FROM AUTHOR]

The purpose of this study is to solve the problem of production planning to minimize overstock with forecasting methods where the results are used to form a Goal Programming model and solved by the Weighted method. Overstock problems can be solved by using the demand forecasting method by comparing the MAD, MSE, and MFE values of four forecasting methods including the Double Moving Average, Single Exponential Smoothing, Double Exponential Smoothing, and Linear Trend method. Forecasting results we obtained are then used to form the Goal Programming model. Another goal to be achieved is to reduce inventory costs and also get profit from sales so that the Weighted method is used here because each target has a different importance weight. The amount of the estimated cost of inventory due to forecast in 2020 will decrease from Rp. 8,130,505,919 to around Rp. 4,589,859,000. Sales profit is obtained with an income of Rp. 379,942,200. [ABSTRACT FROM AUTHOR]

Outlier is an observation in a data set that has a different pattern from other observations in the data set that appears in the form of extreme values in time series or panel data. Outliers can be caused by data input errors, sampling errors, or facts in sample. One method to overcome outliers is the robust method. In time series data, models that are generally used are Exponential Smoothing, Moving Average, and Exponential Weighted Moving Average (EWMA). One model that is often used is the EWMA model. The standard EWMA model is not sensitive to data that has outliers, so the EWMA model that is robust against outliers is needed. The EWMA model for multivariate time series data is called the Multivariate Exponential Weighted Moving Average (MEWMA) method, but this model is also prone to outliers. This is caused by the use of classical mean vector and covariance matrix. To solve this problem, a Modified One-Step M-Estimator (MOM) is used. MOM is a efficient robust estimator for multivariate models. The purpose of this study is to estimate the parameter model with MOM in order to obtain a robust MEWMA model that produce better estimates for multivariate data that has outliers. [ABSTRACT FROM AUTHOR]

We consider the extremal problem of interpolation of scattered data in ℝ3 by smooth curve networks with minimal Lp-norm of the second derivative for 1 < p ≤ ∞. The problem for p = 2 was set and solved by Nielson [1]. Andersson et al. [2] gave a new proof of Nielson's result by using a different approach. Vlachkova [3] extended the results in [2] and solved the problem for 1 < p < ∞. The minimum Lp-norm network for 1 < p < ∞ is obtained from the solution to a system of nonlinear equations with coefficients determined by the data. The solution in the case 1 < p < ∞ is unique. We denote the corresponding minimum Lp-norm network by Fp. In the case where p = ∞ we establish the existence of a solution of the same type as in the case where 1 < p < ∞. This solution on each edge of the underlying triangulation is a quadratic spline function with at most one knot. We denote this solution by F∞ and prove that the minimum Lp-norm networks Fp converge to the minimum L∞-norm network F∞ as p → ∞. [ABSTRACT FROM AUTHOR]

The work examines the monthly dynamics of exports, imports and trade turnover between Ukraine and Kazakhstan from 2005 to 2021. Kazakhstan was one of the largest trading partners of Ukraine. The annual trade in goods between the countries (until 2014) reached about $ 5 billion. Then the military-political situation changed dramatically, as a result of which the volume of trade between the countries fell significantly. Therefore, for modeling and forecasting, we used the data of the monthly dynamics of commodity turnover from January 2015 to the present. Models were used for modeling and forecasting are the Box-Jenkins type: additive models with season component ARIMA (Autoregressive Integrated Moving Average)*ARIMAS (ARIMA for seasonal components or SARIMA) and Holts exponential smoothing. Forecasting of exports, imports and trade turnover for 2021 has been completed. The forecast results showed a small relative error compared to the actual data for July, August and September 2021. [ABSTRACT FROM AUTHOR]

This paper presents a hybrid model between MODWT and FFNN. FFNN can overcome the nonlinear data modeling which is not suitable if it is modeled by ARMA modeling or other models that are actually intended for linear data. MODWT is used in the preprocessing stage. MODWT is able to make data more stationary because it divides the data into detailed and smooth subsets where detailed signals are stationary. If data is stationary, data will be easier to analyze. Each signal (both detail and smooth) is modeled with the FFNN. Then the MODWT-FFNN modeling results are the sum of the FFNN detailed signal and the FFNN smooth signal. Modelling prediction of train passengers in Java excluding Jabodetabek, Indonesia using MODWT-FFNN results forecasting with RMSE is 136.18, while RMSE of ARIMA and FFNN are 560.17 and 234. These results indicate that the MODWT-FFNN model gives a more accurate forecast than ARIMA and FFNN in forecasting train's passengers in the Java. [ABSTRACT FROM AUTHOR]

In this and upcoming age of analytics and decision science, forecasting time series from diversified sources is essentially an important but challenging research area. Although, there are many methods encompassing benchmark approaches to serve the purpose, further robust methodical advances are in escalated demand due to limitations of existing ones. In quest of finding better forecasting approaches, this paper focused on implementation of empirical mode decomposition (EMD) with double exponential smoothing (DES) method (EMD-DES) and performance evaluation of this hybrid time series data fitting and forecasting approach. Both EMD and DES being very effective tools on their area of locally adaptive decomposition and fitting smoothly as well as forecasting respectively, their synergy could be a better forecasting approach as per our intuition. To evaluate EMD-DES method, 06 (six) daily stock price data sets from Nifty 50 index of Indian National Stock Exchange (NSE) were employed and accuracy of this method was compared with other 05(five) benchmark and hybrid methods through 03(three) error measures in 04(four) forecast horizons. These empirical results showed the superiority of EMD-DES method over others. Thus, the method appeals for further attention as a potentially rigorous method. [ABSTRACT FROM AUTHOR]

Coronavirus disease of 2019 (COVID-19) has become widespread within few months and it has lead to a dramatic loss for human life worldwide. This pandemic impacts tens of millions of deaths each day and the number of people were dead by covid-19 is gradually increasing throughout the globe. During this pandemic situation control, we tend to propose a future prediction using Machine Learning algorithms on the death rate, the number of recovered estimates and the number of daily confirmed COVID-19 cases reported within the next ten days. It is based on Machine Learning technique. This forecasting method will predict the upcoming number of COVID-19 cases. Here we use four standard models for forecasting includes linear Regression (LR), The Lowest Absolute and Selective Shrinking Operator (LASSO), Vector Assistance (SVM) and exponential smoothing (ES) will predict the number of COVID-19 cases in future. These four models make three predictions: the mortality rates, the number of newly affected COVID-19 cases and the cummulative number of recovered cases in the next 10 days. These methods are better used in the COVID-19 situation. Based upon the findings, it is a encouraging method to use these standard models in the current situation of COVID-19 spread. The analysis shows that among all the standard forecasting models, ES model performs best, then LR and LASSO which also performs well in predicting the new infected cases of corona, death rate and recovery cases. Whereas the results of SVM were very bad in all the prediction scenarios from the given covid-19 data set. The predictions made by these models relating to the current situation are accurate and will also be useful for future awareness of the future situation. This paper will be enhanced continuously and next we are planning to traverse the prediction methodology using the updated covid-19 data set and we will make use of the most precise and best Machine Learning models for forecasting in future. [ABSTRACT FROM AUTHOR]

The safety and availability of water are important for public health, domestic use, food, and drink production process. Since water is essential in daily life, the demand for water intake and water supply are increasing. Moreover, it must be ensured that the water intake is sufficient to supply water towards the consumers. This is vital to avoid water scarcity in society. Thus, the main objective of this research is to forecast the water intake and water supply by adopting exponential smoothing and Box-Jenkins methods. Then, the forecast performance is evaluated by using mean error (ME), mean absolute scaled error (MASE) and mean absolute percentage error (MAPE). Further, the best model is used to forecast water intake and water supply for the next seven days. The daily data for three years of water intake and water supply in one of the water plants in Johor was collected. The forecast capabilities of the two different methods were compared. Both methods are fitted well but, in overall, the triple exponential smoothing method is outperformed compared to the Box-Jenkins method. This is due to the exponential smoothing method to produce less MAPE and ME values. Both datasets shared the same model which imply the water treatment system used is stable and in good condition. Besides, the water intake and water supply by using triple exponential smoothing method is predicted to be decreased in the following seven days. [ABSTRACT FROM AUTHOR]

Currency exchange rate is one of the external factors that will affect the financial status of Malaysia. Therefore, forecasting the foreign currency exchange rate is important for the financial decision makers, bankers, academic researchers and business practitioners. Time series method is an important area of predicting future data based on the past data. In this study, Auto-Regressive Integrated Moving Average (ARIMA), Double Exponential Smoothing method and Holt-Winter additive method will be used to forecast the data of currency exchange rate of Malaysia Ringgit (RM) to United States of America Dollar (USD). The Mean Absolute Percentage Error (MAPE) for ARIMA, Double Exponential Smoothing method and Holt-Winter additive method are 0.9400, 0.9035 and 2.2686 respectively. In conclusion, the model generated by using Double exponential Smoothing method is the best model to forecast the currency data with the lowest value of MAPE, Mean Absolute Error (MAE) and Mean Square Error (MSE) compared to ARIMA method and Holt-Winter Additive method. [ABSTRACT FROM AUTHOR]

This study provides a time series analysis and interpretation of the output for forecast sales of chicken based food product of weekly sales data. These data were collected directly from the outlet shop of one factory in Malacca started from January 2015 to December 2016. Methods of forecasting include autoregressive (AR) method and simple exponential smoothing (SES) method. The accuracy for both methods will be compared using mean squared error (MSE), mean absolute percentage error (MAPE) and mean absolute deviation (MAD). There will be 1 period ahead of predictions for AR method and 1 period ahead for SES method. This analysis found that AR method with AR (1) model is more accurate than SES method and can be used for the future prediction of chicken based food product of weekly sales data. Recommendations for future study is trying out other method to analyse this sales of chicken based food product and using R software to analyse the dataset. [ABSTRACT FROM AUTHOR]

A noise that recorded at magnetotelluric acquisition data makes the data quality not good enough, so the information obtained after data processing might not be correct or not suitable for the subsurface condition. Several characters of noisy magnetotelluric data are the spiky shaped and non-stationarity time-series curves. This non-stationarity character can't be handled by the Fourier Transformation process. This research used Empirical Mode Decomposition (EMD) in the original of Huang as a filtering method in order to overcome the non-stationarity. This method decomposed the signal into a group of oscillation mode called Intrinsic Mode Decomposition (IMF). It is one of the best IMF chosen as the filtering result by spectrum analysis in the frequency domain. This work used magnetotelluric data from a station that had three components with different frequency sampling, which was 15 Hz, 150 Hz, and 2400 Hz. IMF filtering method then applied to the data resulting in a smoother times-series curve with the suppressed non-stationarity character. This research showed that EMD filtering can be implemented at magnetotelluric data processing and emphasized the effect caused by noise. [ABSTRACT FROM AUTHOR]

Arjuno-Welirang Mountain is located in East Java province, which is a type of stratovolcano with an altitude of 3,339 m for Arjuno and 3,156 m for Welirang. The Arjuno-Welirang geothermal system is a volcanic geothermal system. Magnetotelluric is the most commonly used geothermal exploration method because of its deep penetration and sensitivity to conductive zones. The Magnetotelluric method measures 103 MT points in the field of geothermal systems in the South of Mount Arjuno-Welirang. In this study, the 3D inversion model of Magnetotelluric data was used to reduce the ambiguity of 2D Magnetotelluric inversion data. 3D Model Inversion uses WinGlink and 3DModEM. The steps of Magnetotelluric processing consist of time-domain conversion to the frequency domain, cross power selection, data smoothing, static effect correction, 2D and static strips, interpolation, masking, 3D inversion, and interpretation of inversion models. The 3D model inversion consists of 3D mesh grid parameters used by the NonLinear Conjugate Gradient Algorithm. 2D Inversion Modelling was done by the NLCG method, which produces a 2D model with an RMS error of 1.7%. 3D Inversion modeling was done by the NLCG method producing a 2D model with a value of RMS error of 2.83%. The geothermal system of the Arjuno-Welirang geothermal region consists of a layer of veil rock with a type of resistivity value of 5-11 Ωm, a middle resistivity zone with a value range of 11-76 m acting as a reservoir, and a high resistivity zone with a range of 76-500 Ωm or commonly called heat source. The results of the 3D inversion model indicated a range of low resistance types (hood rock layers) that have a thickness of 1200 - 1700 m, the value of this type of resistance is below the surface with a thickness of ±500-600 m. Thus, the Mt. Arjuno-Welirang is a geothermal prospect area. [ABSTRACT FROM AUTHOR]

We carried out a statistical analysis of the monthly cerebral malaria cases in Papua New Guinea for the period January 1987 - December 1996 and also the analysis of the monthly solar activity and geomagnetic activity for the same period. The solar activity was described by the sunspot numbers and the F10.7 solar flux. As a measure of the geomagnetic activity the aa-indices were used. A close negative correlation between cerebral malaria cases and solar activity indices was found. The correlation of cerebral malaria cases with F10.7 reaches −0.66, and −0.70 with sunspot numbers. The confidence level of relationships was determined by Student’s t-test and reaches 0.00001. All calculations were performed with raw data without smoothing which supports the correctness of the obtained results. The close correlations of cerebral malaria found in Papua New Guinea with solar activity indices could be important for the prognosis of cerebral malaria in this region. [ABSTRACT FROM AUTHOR]

This study deals with forecasting economic time series that have strong trends and seasonal patterns. How to best model and forecast these patterns has been a long-standing issue of time series analysis. In this work, we propose a Holt-Winters Exponential Smoothing approach to time series forecasting in order to increase the chance of capturing different patterns in the data and thus improve forecasting performance. Therefore, the main propose of this study is to compare the accuracy of Holt-Winters models (additive and multiplicative) for forecasting and to bring new insights about the methods used via this approach. These methods are chosen because of their ability to model trend and seasonal fluctuations present in economic data. The models are fitted to time series of e-commerce retail sales in Portugal. Finally, a comparison is made and discussed. [ABSTRACT FROM AUTHOR]

We consider the extremal problem of interpolation of scattered data in R³ by smooth curve networks with minimal Lp-norm of the second derivative for 1 < p < ∞. The problem for p = 2 was set and solved by Nielson [7]. Andersson et al. [1] gave a new proof of Nielson’s result by using a different approach. It allowed them to set and solve the constrained extremal problem of interpolation of convex scattered data in R³ by minimum L²-norm networks that are convex along the edges of an associated triangulation. Partial results for the unconstrained and the constrained problems were announced without proof in [8]. The unconstrained problem for 1 < p < ∞ was fully solved in [10]. Here we present complete characterization of the solution to the constrained problem for 1 < p < ∞. [ABSTRACT FROM AUTHOR]