Smaller MSE Research Articles

B-112 Regression Analysis Techniques for Assessing Calibration Reproducibility Within and Between Instruments. A Practical Example with Blood Ca²⁺ Assayed with Two cobas® b 221Analyzers

Abstract Background Reproducibility of the calibration curve within the analytical measurement range (AMR) is crucial for ensuring reliable patient values within and between laboratory instruments to detect shifts and trends of clinical significance. In this practical example for ionized calcium (Ca²⁺) assayed with two cobas b221® instruments, the equality of calibration lines and the verification with patient specimen comparison study was assessed with regression analysis techniques. Methods Instruments: Two cobas b 221 (Roche Diagnostics). Calibration material: COOX/MSS™ (Five levels, Roche Diagnostics). This was independently assayed either three or five times in parallel with both instruments. The patient specimens were obtained by venipuncture in Na/Li-heparin BD™ tubes (Becton-Dickinson). Upon receipt, specimens were assayed with both instruments within two minutes. The data were transferred to Minitab® (Version 21, Minitab Inc) statistical software and analyzed with polynomial ordinary least squares (POLS), orthogonal and locally weighted scatterplot smoother (lowess) regression analysis models. The aptness of the model was assessed with standardized residuals (std. res.) diagnostics. For acceptable total error the CLIA’s criterion target value ± 10% was adopted. Results While the POLS model did not show probabilistically significant differences between instruments (P=0.8), it showed probabilistically significant differences between years (P=0.003). However, this was due to a small MSE (= 0.003 mg/dL), which is not clinically significant. The test for linearity showed probabilistically significant lack-of-fit (P<0.0001) due to a small MS pure error (<0.0001 mg/dL). The lowess for the standardized deleted residuals by the fitted value showed a quasi-linear relationship. The multiple comparisons (P=0.3) and the Levene’s (P=0.4) tests did not detect probabilistically significant heteroscedasticity of the variance by reference value. The plot of the std. res. by the leverage showed two possible outliers (std. res.>|3|) and no influential observations (Hi<0.5). These diagnostics proved the aptness of the POLS model. The harmonization between the two instruments was verified with patient specimens. The orthogonal regression model showed in the interval 3-6 mg/dL a quasi-linear relationship between the values as obtained with the two instruments. The orthogonal regression line (y=0.1+1.02 n) and the OLS regression line (y=0.03+601 n) were very similar. There was no probabilistically significant lack-of-fit (P=0.5). The plot of the std. res. by the leverage showed one possible outlier (std.res.>|3|) and no influential observations. Additionally, the plot of the relative differences between the values as obtained with b221#9371 and b221#24796 by the values as obtained with b221#9371 showed only one value exceeding half of the total allowable error. Conclusions The POLS model clearly showed that the calibration lines for the two instruments were quasi-linear in the interval 2-10 mg/dL and reproducible within and between instruments. Additionally, this reliable performance was verified with patient specimens in the interval 3-6 mg/dL. The differences between values as obtained with each instrument were within half of the CLIA’s total allowable error. This clearly showed that equality of calibration lines ensured interchangeable patient values between instruments. Finally, electronic transfer of data and the availability of statistical software like Minitab was crucial for conducting this study.

MSE superiority of the unrestricted Stein-rule estimator in a regression model with a possible structural break

This paper investigates the estimation of a linear regression model with a possible structural break at a known point. We analytically derive the exact formulae of the MSE for the restricted SR and PSR estimators, which shrinks the OLS estimator toward the restriction of no structural break. We compare the MSE performance of restricted/unrestricted SR, PSR, and least squared estimators. We analytically show that the unrestricted SR estimator can have a smaller MSE than the restricted SR estimator even when the restriction is correct. Further, our numerical results show that the unrestricted PSR estimator has the best MSE performance over a wide region of parameter space. These results indicate that the use of the unrestricted PSR estimator is recommended even when a structural break may not exist, although the unrestricted PSR estimator does not take the possibility of no structural break into consideration.

Hiding scrambled text messages in speech signals using a lightweight hyperchaotic map and conditional LSB mechanism.

This study presents a lightweight, secure audio steganography system for hiding text messages for transmission over the Internet, with the aim of addressing the current problems of high computational cost and insufficient security identified in earlier studies. We propose a two-phase functioning mechanism. Text characters are first transformed into ASCII code and stored in a vector, which is then divided into three sub-vectors. These sub-vectors are scrambled using two low-complexity operations, namely a forward-backward reading technique and an odd-even index. Two scrambling loops are performed, the first on the small sub-vectors the second on the vector as a whole. In the hiding phase, the speech signal samples are divided into 256 blocks using only 200 values per block, and low-complexity quadratic and the Hénon maps are used to hide the speech signal in a random manner. The conditional LSB is applied as a low-complexity algorithm to identify hidden bits, and a special hyperchaotic map algorithm is developed to randomly choose locations. The proposed approach provides good security for a scrambled text message, with high SNR and PSNR, small MSE and PESQ, a SSIM value of close to one (As indicated in Tables 1, 2, 3, and 4), a BER value of close to zero (as shown in table 8), NCC value near +1 (as shown in table 8), and an MOS value of near five (as described in table 6), as well as a low computational hiding cost.

COMPARISON OF B-SPLINE AND TRUNCATED SPLINE REGRESSION MODELS FOR TEMPERATURE FORECAST

The spline regression model is a nonparametric model and it is applied to data that do not have a certain curve shape and do not have information about it. In this study, the results of the analysis of the B-Spline regression model and the Spline Truncated model were compared on temperature data at several stations on Java Island to obtain the best model that can be used to forecast the temperature for the next few days. Daily temperature data were obtained from BMKG at Semarang, Juanda, Serang, Sleman, Bandung, and Kemayoran stations. The temperature data were modeled with the B-Spline and Spline Truncated regression using the optimal knot point of the GCV, and the best model was obtained. The analysis shows that the B-Spline regression models are better than the truncated Spline models with a fairly small MSE value and a greater coefficient of determination than the truncated Spline model.

Research on pipeline leakage signal denoising using variational mode decomposition and energy value

In order to separate the effective components and the noise components after variational mode decomposition (VMD) and improve the denoising effect of VMD, a denoising approach combining VMD with energy value (EV) and VMD (VMD-EV-VMD) is proposed. First, VMD decomposes the original signal into K intrinsic mode functions (IMF) and then calculates the EV of the probability density function of each IMF component. According to the changing trend of energy value, the effective components and the noise components are distinguished. Subsequently, each noise component is decomposed by VMD again, the effective components are selected by calculating the correlation coefficient (CC) between the input component of the second VMD and the IMF obtained from the second VMD. The effective components selected by VMD twice are used for signal reconstruction. Simulation experiments and pipeline leakage signal processing show that the method of VMD-EV-VMD achieves a greater SNR than the methods of EEMD, VMD-EV, and VMD-WT. and it has a smaller MSE and MAE, A better denoising effect is achieved and the pipeline leakage characteristics are more accurately preserved. In addition, the accuracy of the SVM classifier is used to verify the effectiveness of the proposed method.

A new approach for estimating variance of a population employing information obtained from a stratified random sampling

In this article, we suggest an enhanced estimator for the estimation of finite population variance using twofold auxiliary variable under stratified random sampling. The numerical expressions for the bias and MSE are determined up to the first order of approximation. In order to effectively validate the theoretical findings, three actual data sets are included. Additionally, the application of the suggested estimators is demonstrated using a simulation study. Results of an empirical comparison among the suggested and existing estimators were investigated. To determine how good the suggested estimator, in comparison to the preliminary estimators, the MSE criterion is used. The suggested estimator has a smaller MSE and better PRE than existing estimators, according to numerical results utilizing actual data sets and a simulation analysis.

Elman neural network-based temperature prediction and optimization for lithium-ion batteries

Reliable and precise temperature prediction is one of the most crucial challenges for improving battery performance and preventing thermal runaway. This paper uses a highly adaptive Elman neural network (Elman-NN) to construct a temperature prediction model for lithium-ion batteries in a metal foam aluminum thermal management system. Numerical modeling methods obtain experimental data sets for model training and testing. The input parameters of the neural network prediction model are ambient temperature, battery discharge rate, cooling air flow rate, and state of charge; the output parameters are the maximum, minimum, and average battery temperature. However, due to the limitations of the gradient descent algorithm, the training process of the Elman neural network tends to fall into local optimum solutions. To further improve the prediction accuracy, the Elman-NN structure was optimized using the PSO algorithm, and the model performance was tested and validated. Compared with the original Elman-NN, the hybrid PSO-Elman-NN has smaller MSE and MAE values, with a maximum reduction of 43% and 25%, respectively. For the three test conditions, the maximum predicted temperature difference does not exceed 1.5 K, and the temperature difference decreases further as the discharge time increases. Moreover, the hybrid model’s prediction accuracy is significantly improved, with the coefficients of determination ( R2) increasing by 1.736%, 0.706%, and 1.851%, respectively. The PSO-Elman-NN performed well in terms of compatibility and accuracy of the battery temperature prediction.

APPLICATION OF PANELIZED SPLINE-SPATIAL AUTOREGRESSIVE MODEL TO HIV CASE DATA IN INDONESIA

Spatial regression analysis is a statistical method used to perform modeling by considering spatial effects. Spatial models generally use a parametric approach by assuming a linear relationship between explanatory and response variables. The nonparametric regression method is better suited for data with a nonlinear connection because it does not need linear assumptions. One of the nonparametric regression methods is penalized spline regression (P-Spline). The P-spline has a simple mathematical relationship with mixed linear model. The use of a mixed linear model allows the P-Spline to be combined with other statistical models. PS-SAR is a combination of the P-Spline and the SAR spatial model so that it can analyze spatial data with a semiparametric approach. Based on data from monitoring the development of the HIV situation in 2018, the number of HIV cases in Indonesia shows a clustered pattern that indicate spatial dependence. In addition, the relationship between the number of positive cases and the factors tends to be nonlinear. Therefore, this study aims to apply the PS-SAR model to HIV case data in Indonesia. The resulting model is evaluated based on the estimates of autoregressive spatial coefficient, MSE, MAPE, and Pseudo R2. Based on the results, the PS-SAR model has an autoregressive spatial coefficient similar to the SAR model and has smaller MSE and MAPE than the SAR model.

Improved family of estimators for the population mean using supplementary variables under PPS sampling.

In this article, we suggest an enhanced family of estimators for estimation of population mean employing the supplementary variables under probability proportional to size sampling. Up to the first order of approximation, numerical formulations of the bias and mean square error of estimators are obtained. From our suggested improved family of estimators, we give sixteen different members. The recommended family of estimators has specifically been used to derive the characteristics of sixteen estimators based on the known population parameters of the study as well as auxiliary variables. The performances of the suggested estimators have been assessed using three actual data. Furthermore, a simulation investigation is also accompanied to evaluate the effectiveness of estimators. The proposed estimators have a smaller MSE and an advanced PRE when linked to existing estimators, which are based on actual data sets and simulation studies. Theoretically and empirically studies also reveal that the suggested estimators accomplish well than the usual estimators.

Comparing Edgeworth Expansion and Saddlepoint Approximation in Assessing the Asymptotic Normality Behavior of A Non-Parametric Estimator for Finite Population Total

Sample surveys concern themselves with drawing inferences about the population based on sample statistics. We assess the asymptotic normality behavior of a proposed nonparametric estimator for finite a population total based on Edgeworth expansion and Saddlepoint approximation. Three properties; unbiasedness, efficiency and coverage probability of the proposed estimators are compared. Based on the background of the two techniques, we focus on confidence interval and coverage probabilities. Simulations on three theoretical data variables in R, revealed that Saddlepoint approximation performed better than Edgeworth expansion. Saddlepoint approximation resulted into a smaller MSE, tighter confidence interval length and higher coverage probability compared to Edgeworth Expansion. The two techniques should be improved in estimation of parameters in other sampling schemes like cluster sampling.

THE PARAMETRIC AND NONPARAMETRIC ESTIMATOR IN SEMIPARAMETRIC REGRESSION FOR LONGITUDINAL DATA WITH SPLINE APPROACH

Regression analysis aims to determine the relationship between response variables and predictor variables. There are three approaches to estimate regression curves, there are parametric, nonparametric, and semiparametric regression. In this study, the form of spline semiparametric regression curve estimator for longitudinal data assessed. Based on the estimator that be obtained by using Weighted Least Square (WLS) optimization applied to model electricity consumption in Madura by choosing a model for longitudinal data based on linear spline estimator with two knot. The good criterion of the model is using the GCV value, the coefficient of determination and the value of MSE. The best model is a model that has a high coefficient of determination and a small MSE value. This spline model has a determination coefficient value of 99,72911% and MSE 32,50458.

Crop water use estimation of drip irrigated walnut using ANNs and ANFIS models

Walnut trees, as well as their fruits, represent an important sector of the agricultural industry and their cultivation significantly contributes to the global economy. Irrigation is a key factor in walnut cultivation and the most important problem is related to accurately estimating the need for irrigation water. Walnut water use was estimated in this study through the artificial intelligence methods of Artificial Neural Networks (ANNs) and Adaptive Neuro-Fuzzy Inference System (ANFIS) using meteorological data in western Türkiye, which has semi-arid climatic conditions. Probabilistic scenarios based on maximum, minimum and average temperature, wind speed and sunshine hours over the period 2016-2019 were developed and tested with ANNs and ANFIS models to estimate walnut evapotranspiration. Results indicate that the optimum performance in the training and testing for ANNs and ANFIS models was obtained from the fourth scenario with R = 0.95 and two climate parameters -sunshine duration and mean temperature-. Both ANNs and ANFIS models were able to predict crop water use obtaining high correlation and the minimum number of climatic parameters. Nevertheless, the ANFIS model had a higher predictive capacity, with smaller MSE (0.36 for training and 0.29 for testing) compared to the ANNs model.

Hybrid Extreme Learning Machine dan Firefly Algorithm untuk Meramalkan Nilai Tukar Rupiah terhadap Dolar

Every country has a currency as a medium of exchange and the movement of its exchange rate can affect the economy of the country. In Indonesia, since the freely floating exchange rates system has been applied in August 1997, the value of rupiah currency in the foreign exchange market can change at any time. Considering the massive impacts of exchange rate fluctuation on the economy, then forecasting the exchange rate of rupiah against the US dollar is important to help Indonesia’s economic growth. The aims of this thesis is to predict the estimated exchange rate of rupiah against the US dollar in the future by using hybrid artificial neural network extreme learning machine (ELM) method and firefly algorithm (FA). In the training process, ELM-FA hybrid has a role to obtain the best weight and bias. The weight and bias that obtained will be used for forecasting and to know the success rate of the training process, the validation test process is required. Based on the implementation of program and simulation for some parameter values on the exchange rate data from Jan 2015 until Jan 2018, with four input and hidden nodes, and one output node, obtained the smallest MSE of the training is 0.000480513 with MSE of the testing is 0.0000854107. The relatively small MSE value indicates that ELM-FA network is able to recognize the data pattern well and able to predict the test data well.

Comparing the Principal Regression Analysis Method with Ridge Regression Analysis in Overcoming Multicollinearity on Human Development Index (HDI) Data in Regency/City of East Java in 2018

East Java Province is one of the densely populated provinces. The population density in the province of East Java does not hamper the welfare of its people. The welfare of the people in the province of East Java can be seen through the Human Development Index (HDI), which includes elements of life expectancy, education, and economy. According to the Central Statistics Agency (BPS) of East Java, the HDI level from 2017 to 2018 has increased, but this data experiences a case of multicollinearity. Researchers compared the Principal Regression Analysis method with Ridge Regression Analysis in overcoming the multicollinearity. The conditions taken are methods that have a small MSE value and a large R 2 value. Based on the results and discussion, the Main Component Regression Analysis method has an MSE value of 2.51 and an R 2 value of 91.0%, the Ridge Regression Analysis Method has an MSE value of 0,00071 and an R 2 of 97.7%. The provisions obtained are the value of MSE Regression Ridge is smaller than the value of MSE Regression Main Components and the value of R 2 Regression Ridge is greater than the value of R 2 Main Components. So, the best method in overcoming the multicollinearity problem in the Regency / City Human Development Index (HDI) data in East Java in 2018 is the Ridge Regression Analysis method.

PREDIKSI KADAR PARTICULATE MATTER (PM10) MENGGUNAKAN JARINGAN SYARAF TIRUAN DI KOTA PEKANBARU

This aims of this study is to predict particulate matter (PM10) levels in Pekanbaru using back propagation artificial neural networks (ANN) based on weather factors. The data used in the form of data from 2014 – 2017 as training data and 2018 data as test data. The architecture proposed is composed of 5 – 5 – 1 neurons and uses the logig-logsig-purelin functions. The training process produces a traincgb with a small MSE value and in the process of testing the PM10 prediction compared to BMKG data has an average error of 26.9062%.

On Some Smooth Estimators of the Quantile Function for a Stationary Associated Process

Let {Xn, n ≥ 1} be a sequence of stationary non-negative associated random variables with common marginal distribution function F(x) and quantile function Q(u), where Q(u) is defined as F(Q(u)) = u. Here we consider the smooth estimation of Q(u), adapted from generalized kernel smoothing (Cheng and Parzen J. Stat. Plann. Infer. 59, 291–307, 1997) of the empirical quantile function. Some asymptotic properties of the kernel quantile estimator, for associated sequences, are also established parallel to those in the i.i.d. case. Various estimators in this class of estimators are contrasted, through a simulation study, among themselves and with an indirect smooth quantile estimator obtained by inverting the Poisson weights based estimator of the distribution function studied in Chaubey et al. (Statist. Probab. Lett. 81, 267–276, 2011). The indirect smoothing estimator seems to be the best estimator on account of smaller MSE, however, a quantile estimator based on the Bernstein polynomials and that using the corrected Poisson weights turn out to be almost as good as the inverse distribution function estimator using Poisson weights.

An optimized JPEG-XT-based algorithm for the lossy and lossless compression of 16-bit depth medical image

JPEG-based compression is the most widely used image compression algorithm. Previously, JPEG-XT-based work focused on 16-bit depth high-dynamic satellite infrared images, but not on medical images. In this study, we represent an optimized JPEG-XT method (OPT_JPEG-XT) that better compresses 16-bit depth medical images by amplifying (N times) discrete cosine transform (DCT) coefficients. The results show that the small integers and the first two decimal portions of DCT coefficients play important roles in the compression of medical images. By using the appropriate N and number of decimal portions (NDP), OPT_JPEG-XT could realize lossless compression of medical images. Regarding upper and lower 8-bit subimages, the upper subimages have more important roles in the improvement of OPT_JPEG-XT performance than the lower subimages; lower subimages could occupy over 90% sizes of the entire encoding files. Thus, OPT_JPEG-XT could save about 60% of storage space with high PSNR (peak signal-noise-ratio, over 100) and low MSE (mean-square-error, less than 0.08) by decreasing the compression efficiency of lower subimages. Compared to the conventional JPEG-XT and JPEG 2000, OPT_JPEG-XT can acquire a similar compression performance (PSNR > 100, MSE < 0.9, SSR (saving space rate, >60%) to JPEG 2000 when using N = 20 for lower subimages and lossless compression of upper subimages. OPT_JPEG-XT could obtain high SSR (about 90%, similar to traditional JPEG-XT) with with much smaller MSE (25 times lower than conventional JPEG-XT). Therefore, OPT_JPEG-XT could be developed a novel compression method that could realize the lossless and lossy compression of medical images.

Missing item responses in latent growth analysis: Item response theory versus classical test theory.

In medical research, repeated questionnaire data is often used to measure and model latent variables across time. Through a novel imputation method, a direct comparison is made between latent growth analysis under classical test theory and item response theory, while also including effects of missing item responses. For classical test theory and item response theory, by means of a simulation study the effects of item missingness on latent growth parameter estimates are examined given longitudinal item response data. Several missing data mechanisms and conditions are evaluated in the simulation study. The additional effects of missingness on differences in classical test theory- and item response theory-based latent growth analysis are directly assessed by rescaling the multiple imputations. The multiple imputation method is used to generate latent variable and item scores from the posterior predictive distributions to account for missing item responses in observed multilevel binary response data. It is shown that a multivariate probit model, as a novel imputation model, improves the latent growth analysis, when dealing with missing at random (MAR) in classical test theory. The study also shows that the parameter estimates for the latent growth model using item response theory show less bias and have smaller MSE’s compared to the estimates using classical test theory.

MODELING CENTRAL JAVA INFLATION AND GRDP RATE USING SPLINE TRUNCATED BIRESPON REGRESSION AND BIRESPON LINEAR MODEL

Inflation and Gross Regional Domestic Income (GRDP) are two macroeconomic variables of a region that are correlated with each other. GRDP prices constant (real) can be used as an indicator of economic growth in a region from year to year. Inflation is calculated from the CPI rate and economic growth is calculated from the GRDP rate. Inflation and economic growth in an area are influenced by several factors including bank interest rates. Analysis of data consisting of 2 correlated responses can be performed with birespon regression analysis. In this research, modeling of inflation data and the rate of GRDP through birespon data modeling uses spline truncated nonparametric method and birespon linear parametric method. The purpose of this study is to model inflation data and the Central Java GRDP rate using spline truncated birespon regression. The results are compared with the birespon linear regression model. By using quarterly data from the first quarter of 2007 - the second quarter of 2019, the spline truncated model is better than the linear model, because the spline truncated model has a smaller MSE and R2 is greater than the linear model. Both models have the same performance which is quite good.

Peramalan Rate of Return Saham Menggunakan Metode Brown’s Weighted Exponential Moving Average dengan Optimasi Levenberg-Marquardt

Stocks can be defined as a sign of the participation of unilateral capital in a company or a limited liability company in the form of proof of a company's assets and it is formed as a valueble proof letters as a statement of participating in capital. Return stockis one of the factors that motivates investors to invest or interact and also a reward for the courage of investors to take risks or investments. PT Bank Rakyat Indonesia is one of the largest banks in Indonesia that experiences an increase in stock prices every year. The Brown's Weighted Exponential Moving Average (BWEMA) method which is a combination of the Weighted Exponential Moving Average (WEMA) and Brown's Double Exponential Smoothing (BDES) methods will be used in this study. The data used is the daily data of closing price of the stock (closing price) of PT. Bank Rakyat Indonesia on February 6, 2018 until February 6, 2019. MSE and MAPE BWEMA values ​​were obtained at 6124.222 and 1.831685%, while MSE and MAPE WEMA values ​​were 7559.211 and 1.998439% respectively. The results obtained show that the BWEMA method has smaller MSE and MAPE values. This shows that the BWEMA method is better than the WEMA method in terms of forecasting. As a results the BWEMA method is continued to calculate the forecasts rate of return for the next 7 days. The average value returns obtained is 0.111497% stock which means during the period of the investment, investors get a profit of 0.111497%% per day of the total funds invested in the shares of PT Bank Rakyat Indonesia (Persero) Tbk.