Smaller Mean Square Error Research Articles

PENERAPAN REGRESI SPLINE TRUNCATED DATA LONGITUDINAL DUA VARIABEL PREDIKTOR UNTUK PEMODELAN HARGA SAHAM PERBANKAN

Stock prices movement are influenced by several factors, including inflation and the exchange rate. The bank is one of the institutions that needs to pay attention to those factors. In this study, analysis of factors that affect stock price can be carried out with the spline truncated approach. The advantage of spline is being able to estimate the data pattern and adjusts to its movement. Stock prices tend to increase and decrease at certain sub-intervals, thus these can be applied in spline truncated. Spline truncated best model for longitudinal data with two predictor variables is determined by choosing the order and optimal knots using the smallest MSE (Mean Square Error). This study used monthly data from January 2019 to December 2022 with a comparison of in sample data and out sample data, which is 90%:10%. The results of the analysis showed the best model obtained on the 2nd orde with 3 knot points. The value is 97.49%, meaning that the model is strong and the MAPE value is 12.71%, which is means that the model has good forecasting ability because it is in the category of 10% ≤ MAPE <20%.

Improved Variance Estimation From Trimmed Samples

ABSTRACTSeveral methods for the robust estimation of the variance of a normal random variable are based on trimming. For example, the popular estimator is based on trimming the pairs of samples with the largest absolute distance. Another example is least trimmed squares regression, where all unusually large residuals are removed before the estimation of the variance. Here, in this work, we propose two new classes of estimators that use an optimal linear combination of the trimmed samples to achieve a lower mean squared error (MSE). The first class guarantees the smallest variance among all unbiased estimators that are based on linear combinations, while the second class of estimators guarantees the smallest MSE at the cost of some small bias. Through simulations, we demonstrate that our proposed estimators can have considerably smaller MSE than other robust estimators both in the presence and absence of outliers.

How well do ridge parameter estimators proposed so far perform in terms of normality, outlier detection, and MSE criteria?

Ridge regression is a commonly used prediction method in cases of multicollinearity among regressors in multiple linear regression model. In this study, the performances of 366 different estimators proposed in the literature for ridge parameter estimation were evaluated. The criteria of estimators’ ratio of generating outliers and the proportion of normality besides the Mean Square Error (MSE) criterion were used for diagnosing ridge parameter estimators. Besides, the effect of outliers and deviations from normality generated by the ridge parameter estimators on the MSE was revealed by Spearman’s Rho correlation analysis. To do so, a Monte Carlo simulation was designed for 2420 different cases, considering the simulation parameters comprehensively in terms of the number of regressors, sample size, collinearity level and error variances. The results revealed that some ridge parameter estimators recommended according to MSE criterion, do not perform well in terms of outlier generation and normality criteria. Significant inter-linear correlations were also found among the ridge parameter estimators’ MSE performance criterion, ratios of outlier generation criterion and normality criterion. Thus, it is suggested that an effective ridge parameter estimator should perform well in terms of outlier generation and normality criteria, as well as having small MSE value.

A new method of measuring Forbush decreases

Context. Forbush decreases (FDs) are short-term depressions in the galactic cosmic ray (GCR) flux and one of the common signatures of coronal mass ejections (CMEs) in the heliosphere. They often show a two-step profile, the second one associated with the CME’s magnetic structure. This second step can be described by the recently developed analytical FD model for flux ropes (FRs) – ForbMod. Aims. The aim of this study is to utilise ForbMod to develop a best-fit procedure to be applied on FR-related FDs as a convenient measurement tool. Our motivation is to develop a best-fit procedure that can be applied to a data series from an arbitrary detector. Thus, the basic procedure would facilitate measurement estimation of the magnitude of the FR-related FD, with the possibility of being adapted for the energy response of a specific detector for a more advanced analysis. Methods. The non-linear fitting was performed by calculating all possible ForbMod curves constrained within the FR borders to the designated dataset and minimising the mean square error (MSE). In order to evaluate the performance of the ForbMod best-fit procedure, we used synthetic measurements produced by calculating the theoretical ForbMod curve for a specific example CME and then applying various effects to the data to mimic the imperfection of the real measurements. We also tested the ForbMod best-fit function on the real data, measured by detector F of the SOHO/EPHIN instrument on a sample containing 30 events, all of which have a distinct FD corresponding to the magnetic obstacle. The extraction of FD profiles (from the onset to the end) was performed manually by an observer, whereby we applied two different versions of border selection and assigned a quality index to each event. Results. We do not find notable differences between events marked by a different quality index. For events with a selection of two different borders, we find that the best fit applied on extended interplanetary coronal mass ejection (ICME) structure borders results in a slightly larger MSE and differences compared to the traditional method due to a larger scatter of the data points. We find that the best-fit results can visually be categorised into six different FD profile types. Although some profiles do not show a visually pleasing FD, the ForbMod best-fit function still manages to find a solution with a relatively small MSE. Conclusions. Overall, we find that the ForbMod best-fit procedure performs similar to the traditional algorithm-based observational method, but with slightly smaller values for the FD amplitude, as it’s taking into account the noise in the data. Furthermore, we find that the best-fit procedure has an advantage compared to the traditional method as it can estimate the FD amplitude even when there is a data gap at the onset of the FD.

Application of ARIMA (Autoregressive Integrated Moving Average) model to predict Rupiah selling exchange rate against US Dollar

Currency is a tool in the form of money that is accepted and valid and legal as a payment and economic transactions in a country. US dollar is benchmark for world currencies, so predicting rupiah against US dollar is important. The purpose of this study is to analyze the characteristics of daily selling rate Rupiah against US Dollar, determine best model, and make predictions selling rate of Rupiah against US Dollar. Data used is daily data on selling rate of Rupiah against US Dollar 20 November 2020 - 19 January 2023 with details data training 20 November 2020 - 20 November 2022 and data testing 21 November 2022 - 19 January 2023. The model used is Autoregressive Integrated Moving Averages (ARIMA). The best model was chosen based on the smallest Mean Squared Error (MSE), Mean Absolute Percentage Error (MAPE) and Akaike Information Criterion (AIC). From the analysis results, it is found that best model is ARIMA (2,1,2) because it has significant parameters, white-noise residuals and has smallest MSE and MAPE values. With ARIMA model (2,1,2) the forecasting results for January 20 2023 – January 31 2023 is obtained with highest selling price on January 30 2023 Rp.15,932.4 and smallest on January 20 2023 Rp.15,901.9 and the average Rp.15,919.4. Based on these results, exporters and importers can consider their business activities.

Developing ridge estimators for the extended Poisson-Tweedie regression model: Method, simulation, and application

The extended Poisson-Tweedie (EPT) regression model is one of the count data regression models. It's a more flexible model in count data; since EPT model can be used in two situations: over-dispersion or under-dispersion but under assumption that this model uses second moment to be more flexible model. However, when the predictor (explanatory) variables of the EPT model are highly correlated, this means that there is the multicollinearity problem in the model. It causes inflation of the standard error of the maximum likelihood estimates and makes some of the significant variables insignificant. To handle and reduce the impact of the multicollinearity problem in the EPT model, we developed ridge estimators for this model. A theoretical comparison between the maximum likelihood and the proposed ridge estimators is done. The efficiency of the estimators is evaluated using the mean squared error (MSE). The simulation study and real-life application are used to evaluate the proposed estimators. We examined the performance of seven ridge estimators of the biasing parameter (k) to determine the most appropriate ridge estimator for this model. The simulation and application results showed that the proposed ridge estimator is superior to the maximum likelihood estimator, as it has the smallest MSE.

Temporal data-driven short-term traffic prediction: Application and analysis of LSTM model

With the development of the economy, the number of private cars has increased, and traffic congestion is becoming increasingly common. In order to prevent traffic congestion, short-term prediction of traffic flow is urgent. This article is based on the data of a Hangzhou elevated bridge from October 4th to 18th, 2015, with a timestamp of 5 minutes and a total of over 4000 pieces of data, to have the Long Short-Term Memory (LSTM) model to be trained and the following days traffic flow to be predicted. The outcomes show that the model has high predictive capability and can reliably forecast future short-term transportation passenger flow, accurately reflecting the trend of data changes. Compared with the Auto-Regressive Moving Average (ARIMA) method with a Mean-Square Error (MSE) of 36, the LSTM model has a MSE of 22, which indicates a smaller MSE, indicating a more accurate prediction performance of LSTM. The LSTM models prediction is beneficial for alleviating traffic congestion, providing great convenience for peoples daily travel, and greatly reducing their travel time.

Spearman-like correlation measure adjusting for covariates in bivariate survival data.

We propose an extension of Spearman's correlation for censored continuous and discrete data that permits covariate adjustment. Previously proposed nonparametric and semiparametric Spearman's correlation estimators require either nonparametric estimation of the bivariate survival surface or parametric assumptions about the dependence structure. In practice, nonparametric estimation of the bivariate survival surface is difficult, and parametric assumptions about the correlation structure may not be satisfied. Therefore, we propose a method that requires neither and uses only the marginal survival distributions. Our method estimates the correlation of probability-scale residuals, which has been shown to equal Spearman's correlation when there is no censoring. Because this method relies only on marginal distributions, it tends to be less variable than the previously suggested nonparametric estimators, and the confidence intervals are easily constructed. Although under censoring, it is biased for Spearman's correlation as our simulations show, it performs well under moderate censoring, with a smaller mean squared error than nonparametric approaches. We also extend it to partial (adjusted), conditional, and partial-conditional correlation, which makes it particularly relevant for practical applications. We apply our method to estimate the correlation between time to viral failure and time to regimen change in a multisite cohort of persons living with HIV in LatinAmerica.

Impact of nausea/vomiting on EQ-5D-5L utility scores in patients taking iron preparations for heavy menstrual bleeding or anemia

BackgroundThe purpose of this study was to establish an estimating equation to predict the 5-level EQ-5D version (EQ-5D-5L) utility score in patients taking iron preparations for heavy menstrual bleeding (HMB) or anemia and to evaluate whether the presence of nausea or vomiting was a significant predictor of EQ-5D-5L-based quality of life.MethodsA cross-sectional survey was conducted to collect EQ-5D-5L utility scores and other patient reported outcomes from 385 patients taking iron preparations for HMB or anemia who were selected from the disease patient panel. Using the utility scores as objective variables, explanatory variable candidates were selected considering correlations, multicollinearity, and clinical validity. Predicting models were constructed using regression-based models (linear model, generalized linear model (GLM), Tobit model). Stepwise regression method was applied for selecting statistically significant (p < 0.05) predictors. Goodness-of-fit of models were assessed by mean absolute error and mean squared error (MSE).ResultsThe EQ-5D-5L utility scores (mean ± standard deviation) of 96 patients with nausea/vomiting and 289 patients without nausea/vomiting were 0.67 ± 0.16 and 0.84 ± 0.14, respectively (p < 0.001). The presence of nausea/vomiting was shown to be the most significant factor reducing the utility score in the statistical models using the explanatory variable candidates selected in the study. As the results of the goodness-of-fit test, GLM with the smallest MSE was selected to establish the estimating equation.ConclusionThe estimating equation to predict the EQ-5D-5L utility scores in patients taking iron preparations for HMB or anemia was established. The presence of nausea/vomiting was found to be a factor significantly reducing utility scores, with a decrement of the value estimated to be -0.117.Trial registrationUMIN000045700 (http://www.umin.ac.jp/ctr/). Registered on October 11, 2021.

Evaluation of Bayesian spatiotemporal infectious disease models for prospective surveillance analysis

BackgroundCOVID-19 brought enormous challenges to public health surveillance and underscored the importance of developing and maintaining robust systems for accurate surveillance. As public health data collection efforts expand, there is a critical need for infectious disease modeling researchers to continue to develop prospective surveillance metrics and statistical models to accommodate the modeling of large disease counts and variability. This paper evaluated different likelihoods for the disease count model and various spatiotemporal mean models for prospective surveillance.MethodsWe evaluated Bayesian spatiotemporal models, which are the foundation for model-based infectious disease surveillance metrics. Bayesian spatiotemporal mean models based on the Poisson and the negative binomial likelihoods were evaluated with the different lengths of past data usage. We compared their goodness of fit and short-term prediction performance with both simulated epidemic data and real data from the COVID-19 pandemic.ResultsThe simulation results show that the negative binomial likelihood-based models show better goodness of fit results than Poisson likelihood-based models as deemed by smaller deviance information criteria (DIC) values. However, Poisson models yield smaller mean square error (MSE) and mean absolute one-step prediction error (MAOSPE) results when we use a shorter length of the past data such as 7 and 3 time periods. Real COVID-19 data analysis of New Jersey and South Carolina shows similar results for the goodness of fit and short-term prediction results. Negative binomial-based mean models showed better performance when we used the past data of 52 time periods. Poisson-based mean models showed comparable goodness of fit performance and smaller MSE and MAOSPE results when we used the past data of 7 and 3 time periods.ConclusionWe evaluate these models and provide future infectious disease outbreak modeling guidelines for Bayesian spatiotemporal analysis. Our choice of the likelihood and spatiotemporal mean models was influenced by both historical data length and variability. With a longer length of past data usage and more over-dispersed data, the negative binomial likelihood shows a better model fit than the Poisson likelihood. However, as we use a shorter length of the past data for our surveillance analysis, the difference between the Poisson and the negative binomial models becomes smaller. In this case, the Poisson likelihood shows robust posterior mean estimate and short-term prediction results.

Modeling of Knowledge Toward Herbal Medicine for Oral Health Using Multiple Linear Regression and Neural Network.

Background and goals Herbal medicine is used to treat a variety of oral health problems. Therefore, it is essential to comprehend it fully. To determine whether the amount used is risky, it is crucial to understand the dosages of medicinal plants. Before performing multiple linear regression (MLR) modeling, this paper uses the multilayer feedforward (MLFF) neural network (NN) technique to propose the variable selection. A data set with socio-demographic variables for dental staff and herbal medicine related to oral health knowledge score (KS) was chosen to demonstrate the design-build methodology. Materials and methods It was discovered that the KS is significantly related to the sex, age, income, occupation, and practice score (PS) at the first stage of the selection process, where all the variables were screened for their clinical importance. These five variables are chosen and used as inputs for the MLFF model by considering the level of significance, alpha = 0.05. Then, using the best variable discovered by the MLFF process, the MLR is applied. Results The performance of MLFF was evaluated using the mean squared error (MSE). MSE measures how far our estimates are off from the actual results. The MLFF's smallest MSE indicates the model's ideal combination of variable selection. Conclusion This study showed that using MLFF would help confirm the selected independent variables for MLR. In addition, KS level is more correlated with occupation, PS, and sex than with age and income. Moreover, this model could be used practically forany dataset.with the same criteria.

Infilling missing data and outliers for a conventional sewage treatment plant using a self-organizing map: a case study of Kauma Sewage Treatment Plant in Lilongwe, Malawi

Abstract Data availability is key for modeling of wastewater treatment processes. However, process data are characterized by missing values and outliers. This study applied a self-organizing map (SOM) to fill in missing values and replace outliers in wastewater treatment data from Kauma Sewage Treatment Plant in Lilongwe, Malawi. We used primary and secondary wastewater data and executed the SOM algorithm to fill missing values and replace outliers in effluent pH, biochemical oxygen demand, and dissolved oxygen. The results suggest that the SOM algorithm is reliable in filling gaps in wastewater time series data with less than 50% missing values with correlation coefficient (R) values of &amp;gt;0.90. The SOM algorithm failed to reliably fill gaps and replace outliers in time series data with &amp;gt;50% missing values. For instance, high mean square error (MSE) values of 3,655.57, 10.62, and 2,153.34 for pH, DO, and BOD, respectively, were registered in datasets with more than 50% missing values, while very small MSE values (MSE ≈ 0) were associated with effluent pH, BOD, and DO data with missing values of &amp;gt;50%. Practitioners can use this approach to improve the planning and management of wastewater treatment facilities where available data records are riddled with missing observations.

Forward and Inverse Kinematics Solution of A 3-DOF Articulated Robotic Manipulator Using Artificial Neural Network

In this research paper, the multilayer feedforward neural network (MLFFNN) is architected and described for solving the forward and inverse kinematics of the 3-DOF articulated robot. When designing the MLFFNN network for forward kinematics, the joints' variables are used as inputs to the network, and the positions and orientations of the robot end-effector are used as outputs. In the case of inverse kinematics, the MLFFNN network is designed using only the positions of the robot end-effector as the inputs, whereas the joints’ variables are the outputs. For both cases, the training of the proposed multilayer network is accomplished by Levenberg Marquardt (LM) method. A sinusoidal type of motion using variable frequencies is commanded to the three joints of the articulated manipulator, and then the data is collected for the training, testing, and validation processes. The experimental simulation results demonstrate that the proposed artificial neural network that is inspired by biological processes is trained very effectively, as indicated by the calculated mean squared error (MSE), which is approximately equal to zero. The resulted in smallest MSE in the case of the forward kinematics is 4.592×10^(-8) in the case of the inverse kinematics, is 9.071×10^(-7). This proves that the proposed MLFFNN artificial network is highly reliable and robust in minimizing error. The proposed method is applied to a 3-DOF manipulator and could be used in more complex types of robots like 6-DOF or 7-DOF robots.

Forecasting The Search Trends of The Keyword “Sarung Wadimor” In Indonesia on Google Trends Data Using Time Series Regression with Calender Variation and Arima Box-Jenkins

The impact of this 4.0 era is that data is growing and can be collected very easily and then reprocessed to obtain information. One of the search engines for various data and information that is often used is Google, causing a high search intensity and will further impact on increasing the amount of data generated by search engines. Google Trends is one of the official websites from Google that reflects or takes pictures of events in society based on search keywords. The search keyword that will be studied in this article is “Sarung Wadimor”. Therefore, the purpose of this research is to forecast the search trend for the keyword "Sarung Wadimor" which is interesting because the resulting time series data pattern shows a recurring pattern due to the effect of calendar variations which are thought to be related to the month of Ramadan. Forecasting modeling uses Autoregressive Integrated Moving Average (ARIMA) and Time Series Regression (TSR). The goodness of the model used in this article is the Mean Square Error (MSE), Root Mean Square Error (RMSE), and Symmetric Mean Absolute Percentage Error (SMAPE). Based on the results of the analysis, using three goodness-of-fit measures shows that the TSR model with the Calendar Variation of Ramadan + Month Periods has smaller MSE, RMSE, and SMAPE values than the other models with goodness-of-fit values of 88.602, 9.413, and 26.950, respectively. Forecasting results for the next 6 periods show that the search trend for the keyword "Sarung Wadimor" tends to decrease, this is because the month of Ramadan is still quite far in 2023.

Penalized maximum likelihood inference under themixture cure model in sparse data.

When a study sample includes a large proportion of long-term survivors, mixture cure (MC) models that separately assess biomarker associations with long-term recurrence-free survival and time to disease recurrence are preferred to proportional-hazards models. However, in samples with few recurrences, standard maximum likelihood can be biased. We extend Firth-type penalized likelihood (FT-PL) developed for bias reduction in the exponential family to the Weibull-logistic MC, using the Jeffreys invariant prior. Via simulation studies based on a motivating cohort study, we compare parameter estimates of the FT-PL method to those by ML, as well as type 1 error (T1E) and power obtained using likelihood ratio statistics. In samples with relatively few events, the Firth-type penalized likelihood estimates (FT-PLEs) have mean bias closer to zero and smaller mean squared error than maximum likelihood estimates (MLEs), and can be obtained in samples where the MLEs are infinite. Under similar T1E rates, FT-PL consistently exhibits higher statistical power than ML in samples with few events. In addition, we compare FT-PL estimation with two other penalization methods (a log-F prior method and a modified Firth-type method) based on the same simulations. Consistent with findings for logistic and Cox regressions, FT-PL under MC regression yields finite estimates under stringent conditions, and better bias-and-variance balance than the other two penalizations. The practicality and strength of FT-PL for MC analysis is illustrated in a cohort study of breast cancer prognosis with long-term follow-up for recurrence-free survival.

Highly Precise and Broadband Full-Stokes Polarimeter Based on a Deep Learning Algorithm

Recent developments in nanophotonics enable the practical realization of ultracompact full-Stokes polarimeters and imaging devices based on the well-defined polarization-sensitive structures of subwavelength size. However, the resonant feature of building blocks and the limited ability of the current processing technology greatly limit their detection accuracy and working waveband. Herein, we experimentally demonstrate a highly precise and broadband full-Stokes polarimeter based on the high-performance dielectric chiral shells (DCSs) and the deep learning (DL) algorithm. The large-area dielectric chiral metasurfaces (DCMs) fabricated by depositing Si on the self-assembled microsphere monolayer own a large transmittance, strong chirality, and anisotropy in the visible waveband, which enable the effective perception of the polarization state of light. The Stokes parameters are successfully detected by measuring several transmittances from DCSs with different rotation angles, i.e., Î, and then analyzing with the pre-established mapping relationship between the Stokes parameters Ŝ and the recorded transmittance Î based on the DL algorithm, i.e., Ŝ = f(Î). In contrast to the traditional polarimeters, the detection accuracy based on the DL approach seems insensitive to the fabrication errors. A mean square error (MSE) of smaller than 0.5% at 760.06 nm was recorded, and an averaged MSE of less than 4% for Stokes parameters over a broad waveband from 400 to 840 nm was obtained, although the optical anisotropy and chirality of DCMs own a strong dispersion feature. The small MSE in a broad waveband, together with the easily accessing DCMs, make our polarimeter as the most potential candidate for future applications, such as the color polarization imaging devices and the ultraviolent polarimeter.

Predictive modeling of surface roughness in hard turning with rotary cutting tool based on multiple regression analysis, artificial neural network, and genetic programing methods

This present paper deals with the results of setting up and evaluating the quality of surface roughness (SR) prediction models through hard turning with self-driven rotary cutting tools for 40X steel shaft with hardness 45HRC. The cutting parameters are considered to establish the SR prediction model include angle tilt of cutting tool axis, depth of cut, feed rate, and cutting speed. The predictive models (PM) are established utilizing Multi-variables Regression Analysis (MRA), Artificial Neural Network (ANN), and Genetic Programming (GP) methods. In this regard, four MRA, four ANN, and three GP structures are considered to select the most suitable model. The criteria to estimate the PM quality include Coefficient of Determination ( R2), Mean Square Error (MSE), and Mean Absolute Percent Error (MAPE). Whereby, two data sets were collected to construct regression models (RM) and to serve verification. That dataset was originated from 63 experiments (Ep) including 54 Ep for establishing PM and 9 extensive experiments (EEp) for testing PM. The prediction criteria of the MRA3 model gave the best results in four MRA models with R2 of 0.99, MSE of 0.042, and MAPE of 8.087%. The ANN1 gives the most reliable assessment criteria in four ANN models and the GP3 model gave the best in three GP models. The results of predicting the SR value between the selected MRA and ANN and GP models will be assessed in detail. Accordingly, the evaluation criteria of the ANN1 model are the best with the smallest MSE (0.032) and MAPE (7.207%). The MRA3 and GP3 have a lower confidence predictive criteria value than the ANN1 model.

Stock Price Prediction Based on Simple Decision Tree Random Forest and XGBoost

Stock price prediction is the process of using a variety of methods and techniques to forecast the future performance of the stock. It is helpful for investors and traders to make decisions before and during trading. In this study, three stock prediction models depending on the decision tree algorithm are developed (i.e., Decision Tree, Random Forest and XGBoost) To be specific, their performances are compared to select the best prediction approach. Both accuracy and mean squared error (MSE) are utilized as metrics to evaluate the performances of models. By comparing the results and the metrics calculated according to the predictions, the random forest model is the best model with the highest accuracy and the smallest MSE. However, the XGBoost was the least accurate since this algorithm is too sensitive to the scale of the input variables. These results offer suggestions for choosing stock price prediction models and shed light on improving the models.

Configure the Top Lion Riley distribution uploaded with the application

The Topp-Leone (TL) family is one of the methods used to generalize probability distributions to make them more flexible in application. This generalization was used to generalize the Exponentiated Rayleigh (ER) distribution, this distribution called Topp Leone Rayleigh (TLR). The statistical properties of the new distribution were studied represented by the r-order moments around zero, median, mode, quartiles, and moment generating function, skewness and kurtosis. The parameters of distribution were estimated by the maximum likelihood (ML) method. The results of the estimation method were applied to data generated with different sample sizes and values for the parameters. With the application of data represented by the lengths of the saturated carbon fibres, the research reached the most important conclusions, including that the scale parameter does not affect the skewness and kurtosis measures. While the additional parameter affected the two measures. The TLER distribution became more flexible than the ER distribution. On the practical side, it was found that the mean square error (MSE) decreases with increasing sample size. When the sample size is n = 25,50,100 and the scale parameter is fixed, and When λ = 0.5,1,2 it was found that the estimators when α = 0.75 are the best for having the smallest MSE. The lengths of carbon fibres follow the TLER distribution.

Prediction of Covid-19 Cases in Central Java using the Autoregressive (AR) Method

Since the beginning of the Covid-19 case in Indonesia in March 2020, more than 6 million confirmed cases had been confirmed. The rapid development of this case can be accessed through the covid19.go.id page. In Central Java province, confirmed cases as of July 6, 2022, reached 628,393 people, with the number of recovered patients reaching 594,783 people and the number of patients dying as many as 33,215 people. With this data, a prediction is needed to help the government anticipate an increase in Covid-19 cases in Central Java Province. This study aims to create a forecasting model using the Autoregressive (AR) method by optimizing the function parameters. Then Mean Squared Error (MSE) to analyze the results of forecasting data errors. The results are the best parameter functions on AR (30) with the smallest MSE. Furthermore, predictions are made from July 1 to August 30, 2022, showing an increase in cases