Classical Error Research Articles

Dual Optical Frequency Comb Neuron: Co‐Developing Hardware and Algorithm

Previous studies on photonic neural network have demonstrated that algorithm can inspire hardware design. This study seeks to demonstrate that hardware can also inspire algorithm design. To further exploit the advantages of photonic analog computing, the authors develop hardware and algorithm simultaneously for photonic convolutional neural networks. Specifically, this work developed an architecture called dual optical frequency comb neuron (DOFCN) enabled by an integrated microcomb to perform cosinusoidal nonlinear activation and vector convolution without temporal or spatial dispersion and large‐scale modulator arrays. Furthermore, DOFCN‐based composite vector convolutional neural networks (CVCNNs), an optical‐electric hybrid model, are proposed to perform classification and regression tests in signal modulation format identification and optical structure inverse design tasks, respectively. The ablation experiments show that under 4‐bit precision limit, the element‐wise activation CVCNN has 14% higher classification accuracy, 76% lower regression residuals, and 100% higher training efficiency than that of the 32‐bit standard convolutional neural network (CNN). DOFCN exhibits impressive spectral information processing ability to facilitate signal‐processing tasks related to optics and electromagnetics.

A simplified model to estimate nonlinear turbulent transport by linear dynamics in plasma turbulence

A simplified model to estimate nonlinear turbulent transport only by linear calculations is proposed, where the turbulent heat diffusivity in tokamak ion temperature gradient(ITG) driven turbulence is reproduced for a wide parameter range including near- and far-marginal ITG stability. The optimal nonlinear functional relation(NFR) between the turbulent diffusivity, the turbulence intensity mathcal{T}, and the zonal-flow intensity mathcal{Z} is determined by means of mathematical optimization methods. Then, an extended modeling for mathcal{T} and mathcal{Z} to incorporate the turbulence suppression effects and the temperature gradient dependence is carried out. The simplified transport model is expressed as a modified nonlinear function composed of the linear growth rate and the linear zonal-flow decay time. Good accuracy and wide applicability of the model are demonstrated, where the regression error of sigma _{textrm{model}} = 0.157 indicates improvement by a factor of about 1/4 in comparison to that in the previous works.

Small area estimation of general finite-population parameters based on grouped data

This paper proposes a new model-based approach to small area estimation of general finite-population parameters based on grouped data or frequency data, often available from sample surveys. Grouped data contains information on frequencies of some pre-specified groups in each area, for example, the numbers of households in the income classes. Thus, grouped data provide more detailed insight into small areas than area-level aggregated data. A direct application of the widely used small area methods, such as the Fay–Herriot model for area-level data and nested error regression model for unit-level data, is not appropriate since they are not designed for grouped data. Our novel method adopts the multinomial likelihood function for the grouped data. In order to connect the group probabilities of the multinomial likelihood and the auxiliary variables within the framework of small area estimation, we introduce the unobserved unit-level quantities of interest. They follow a linear mixed model with random intercepts and dispersions after some transformation. Then the probabilities that a unit belongs to the groups can be derived and are used to construct the likelihood function for the grouped data given the random effects. The unknown model parameters (hyperparameters) are estimated by a newly developed Monte Carlo EM algorithm which uses an efficient importance sampling. The empirical best predicts (empirical Bayes estimates) of small area parameters are calculated by a simple Gibbs sampling algorithm. The numerical performance of the proposed method is illustrated based on the model-based and design-based simulations. In the application to the city-level grouped income data of Japan, we complete the patchy maps of the Gini coefficient as well as mean income across the country.

Younger postnatal age is associated with a lower heart rate on Holter monitoring during the first week of life

To evaluate heart rate (HR), the presence of extrasystoles and other Holter findings among healthy newborns, and to collect data for new normal limits for Holter parameters in newborns. For this cross-sectional study, 70 healthy term newborns were recruited to undergo 24-h Holter monitoring. Linear regression analysis was used in HR analyses. The age-specific limits for HRs were calculated using linear regression analysis coefficients and residuals. The mean (SD) age of the infants was 6.4 (1.7) days during the recording. Each consecutive day of age raised the minimum and mean HR by 3.8 beats per minute (bpm) (95% CI: 2.4, 5.2; P < .001) and 4.0 bpm (95% CI: 2.8, 5.2; P < .001), respectively. Age did not correlate with maximum HR. The lowest calculated limit for minimum HR ranged from 56 bpm (aged 3 days) to 78 bpm (aged 9 days). A small number of atrial extrasystoles and ventricular extrasystoles were observed in 54 (77%) and 28 (40%) recordings, respectively. Short supraventricular or ventricular tachycardias were found in 6 newborns (9%).Conclusion: The present study shows an increase of 20 bpm in both the minimum and mean HRs of healthy term newborns between the 3rd and 9th days of life. Daily reference values for HR could be adopted in the interpretation of HR monitoring results in newborns. A small number of extrasystoles are common in healthy newborns, and isolated short tachycardias may be normal in this age group.What is Known:• The current definition of bradycardia in newborns is 80 beats per minute.• This definition does not fit into the modern clinical setting of continuously monitored newborns, where benign bradycardias are commonly observed.What is New:• A linear and clinically significant increase in heart rate was observed in infants between the ages of 3 and 9 days. • It appears as though lower normal limits for heart rate could be applied to the youngest newborns.

Does Fiscal Deficit Cause Economic Growth? Evidence from Nepal

The fiscal deficit soared tremendously after Nepal endorsed economic liberalization aftermath of the 1990s democracy. The purpose of the paper is to investigate the effects of fiscal deficit in the economic growth of Nepal. Data from 1980 to 2019 were used to estimate the short- and long-run causal relations of fiscal deficit proxied by total government expenditure minus total government revenue as a percent of gross domestic product (GDP) and growth rate of real GDP as a proxy of economic growth along with other control variables. The autoregressive distributed lag (ARDL) error correction regression and vector error correction model (VECM) Granger causality have employed to meet the purpose of the study. The study reveals that there is short-run statistically significant negative relationship between Nepal's economic growth and the fiscal deficit of the previous year. However, both domestic and foreign loan and economic growth have positive and significant relations in short run. In contrast, the results reveal that there is positive but not significant relationship between the fiscal deficit and the economic growth of Nepal in long run. The Granger causality suggests that fiscal deficit does not Granger cause the economic growth of Nepal. However, economic growth does Granger causes the fiscal deficit in Nepal. Thus, the growth-driven fiscal deficit will be more effective in the economy of Nepal. Based on stylized facts, it will be necessary to redirect deficit financing into productive channels in order to enhance the sustainable economic growth.

Prediction of KLCI Index Through Economic LASSO Regression Model and Model Averaging

The Financial Times Stock Exchange (FTSE) Bursa Malaysia KLCI Index is a key component in the development of Malaysia's economic growth and the complexity in terms of identifying the factors that have a substantial impact on the Malaysian stock market has always been a contentious issue. In this study, the macroeconomic factors of exchange rate, interest rate, gold price, consumer price index, money supply M1, M2, and M3, industrial production, and oil price were discussed by using economic LASSO regression and Bayesian Model Averaging (BMA) with monthly average and monthly end time-series data spanning from January 2015 to June 2021, with a total of 78 observations by using the R Studio. The findings demonstrate that month-end data is better suited for stock market prediction than month-average data and that the BMA model is more suitable than the LASSO model, as seen by lower Mean Square Error of Prediction, MSE(P) and Residual Mean Square Error of Prediction, RMSE(P) values. The exchange rate, gold price, and money supply have a negative association with the dependent variables, while the consumer price index has a positive relationship associated with the dependent variables. The consumer price index is the most significant contributing factor, whereas gold price is the least significant. The result depicted that the KLCI index has no significant relationship with the variables interest rate, money supply M2, M1, industrial production index, and oil price. In conclusion, investors could specifically focus on the positive contributor and put lesser attention on improving their portfolio return.

Debiased and thresholded ridge regression for linear models with heteroskedastic and correlated errors

AbstractHigh-dimensional linear models with independent errors have been well-studied. However, statistical inference on a high-dimensional linear model with heteroskedastic, dependent (and possibly nonstationary) errors is still a novel topic. Under such complex assumptions, the paper at hand introduces a debiased and thresholded ridge regression estimator that is consistent, and is able to recover the model sparsity. Moreover, we derive a Gaussian approximation theorem for the estimator, and apply a dependent wild bootstrap algorithm to construct simultaneous confidence interval and hypothesis tests for linear combinations of parameters. Numerical experiments with both real and simulated data show that the proposed estimator has good finite sample performance. Of independent interest is the development of a new class of heteroscedastic, (weakly) dependent, and nonstationary random variables that can be used as a general model for regression errors.

A Dendritic Neuron Model Optimized by Meta-Heuristics with a Power-Law-Distributed Population Interaction Network for Financial Time-Series Forecasting

The famous McCulloch–Pitts neuron model has been criticized for being overly simplistic in the long term. At the same time, the dendritic neuron model (DNM) has been shown to be effective in prediction problems, and it accounts for the nonlinear information-processing capacity of synapses and dendrites. Furthermore, since the classical error back-propagation (BP) algorithm typically experiences problems caused by the overabundance of saddle points and local minima traps, an efficient learning approach for DNMs remains desirable but difficult to implement. In addition to BP, the mainstream DNM-optimization methods include meta-heuristic algorithms (MHAs). However, over the decades, MHAs have developed a large number of different algorithms. How to screen suitable MHAs for optimizing DNMs has become a hot and challenging area of research. In this study, we classify MHAs into different clusters with different population interaction networks (PINs). The performance of DNMs optimized by different clusters of MHAs is tested in the financial time-series-forecasting task. According to the experimental results, the DNM optimized by MHAs with power-law-distributed PINs outperforms the DNM trained based on the BP algorithm.

Important Medicinal and Food Taxa (Orders and Families) in Kenya, Based on Three Quantitative Approaches.

Globally, food and medicinal plants have been documented, but their use patterns are poorly understood. Useful plants are non-random subsets of flora, prioritizing certain taxa. This study evaluates orders and families prioritized for medicine and food in Kenya, using three statistical models: Regression, Binomial, and Bayesian approaches. An extensive literature search was conducted to gather information on indigenous flora, medicinal and food plants. Regression residuals, obtained using LlNEST linear regression function, were used to quantify if taxa had unexpectedly high number of useful species relative to the overall proportion in the flora. Bayesian analysis, performed using BETA.INV function, was used to obtain superior and inferior 95% probability credible intervals for the whole flora and for all taxa. To test for the significance of individual taxa departure from the expected number, binomial analysis using BINOMDIST function was performed to obtain p-values for all taxa. The three models identified 14 positive outlier medicinal orders, all with significant values (p < 0.05). Fabales had the highest (66.16) regression residuals, while Sapindales had the highest (1.1605) R-value. Thirty-eight positive outlier medicinal families were identified; 34 were significant outliers (p < 0.05). Rutaceae (1.6808) had the highest R-value, while Fabaceae had the highest regression residuals (63.2). Sixteen positive outlier food orders were recovered; 13 were significant outliers (p < 0.05). Gentianales (45.27) had the highest regression residuals, while Sapindales (2.3654) had the highest R-value. Forty-two positive outlier food families were recovered by the three models; 30 were significant outliers (p < 0.05). Anacardiaceae (5.163) had the highest R-value, while Fabaceae had the highest (28.72) regression residuals. This study presents important medicinal and food taxa in Kenya, and adds useful data for global comparisons.

On the Pulling Linear Regression and Its Applications in Digital Mammograms

Regression analysis is a statistical approach used to investigate the correlations between variables, especially linear regression, that is a simple but effective approach for analyzing the relationship between a dependent variable and one independent variable. Since it has limitations based on the assumption that the mean of the noise should be zero, there are still some areas where it may be improved. In this article, we introduce a novel data fitting algorithm called the pulling linear regression, which is separated into two types: the line-pulling linear regression and the band-pulling linear regression. The method is developed from linear regression, which can create the regression line from the function that uses noise with various distributions. The result demonstrates that the sequence of sum square errors of the pulling linear regression is convergent. Moreover, we have a numerical example to show that the performance of the proposed algorithm is better than that of linear regression when the mean of the noise is not zero. And the last, we have an application to smooth the boundary of the pectoral muscle in digital mammograms. We found that the regression line of the proposed algorithm can do better than the linear regression when we would like to remove only the muscle part.

Validation of Depression, Anxiety and Stress Scale (DASS-21) among healthcare workers during the outbreak of delta variant of SARS-CoV-2 in Ghana

Background: Today COVID-19 is having a dire effect on the mental and physical health of the general population. Although the long-term psychological effects of COVID-19 remain unknown, studies have shown increasing depression, anxiety and stress among healthcare workers. The aim of the study was to examine the psychometric properties and validation of the Depression, Anxiety, and Stress Scale (DASS-21) and evaluate the level of depression, anxiety, and stress among healthcare workers in a tertiary hospital during the SARS-CoV-2 delta variant outbreak. Methods: Using an online questionnaire, we conducted a cross-sectional study on 1201 healthcare workers in a tertiary hospital. The validation of DASS-21 was performed by examining the factorial structure (construct validity) using confirmatory factor analysis (CFA), internal consistency (reliability), convergent validity, discriminate validity and measurement invariance. Results: Cronbach's Alpha was acceptable for depression (0.88), anxiety (0.81), and stress (0.86). CFA provides support for the three-factor oblique model with the following fit indexes: (Chi-Square χ2/ (degrees of freedom) = 1628.5/(186), p &lt; 0.001), comparative fit index (CFI = 0.923), Tucker-Lewis index (TLI = 0.901), standardized root mean square residual (SRMR = 0.0452) and root mean square error of approximation (RMSEA = 0.07). DASS-21 achieved both configural and metric invariance. Conclusions: Our results indicate that DASS-21 is a reliable and valid self-reporting screening tool for depression, anxiety and stress among healthcare workers. This tool is also invariant across sex, doctors, nurses, and non-clinical healthcare workers. Thus DASS-21 is an essential screening tool to identify healthcare workers at a higher risk of developing work-related mental health disorders.

Predicting main engine power and emissions for container, cargo, and tanker ships with artificial neural network analysis

The most significant aspect of international shipping is sea transportation, and the developments to be made in maritime transport will inspire and predict all other fields. Therefore, determining a ship’s main engine power has great importance in terms of both energy efficiency and environmental factors. The maritime transport and shipping industry has currently begun to understand the importance of artificial intelligence technology. This study uses an artificial neural network (ANN) model to predict the main engine power and pollutant emissions of container, cargo, and tanker ships over 14 parameters: maximum speed, average speed, breadth, year built, ship type, status, length overall (LOA), light displacement, summer displacement, fuel type, deadweight tonnage (DWT), gross tonnage, engine cylinder size, and engine stroke length. In order to provide accurate results, the ANN analysis was trained with data from 3,020 ships, which is quite high compared to the studies in the literature. Many ANN models have been developed and compared to achieve minimal errors and highest accuracy in the results. The regression values, which involve the training, validation, and test values for the different ship types, were obtained as 0.99773 for container ships, 0.98964 for cargo ships, and 0.97755 for tanker ships, with a value of 0.97189 for all ships. The ANN structure was tested using many variations for hidden neuron counts, with the ANN analysis made with 30 neurons obtaining the best results. The ANN analysis results were compared with real values, which showed that very accurate results had been obtained according to the mean squared error (MSE), regression, and mean absolute percentage error (MAPE) results. The MSE value had exceeded 20,000 in the two-input ANN model, but decreased to 0.03, 0.081, and 0.13 with the 14-input model for container, cargo, and tanker ships, respectively. In order to make accurate predictions with maximum precision in the ANN analyses, the study attempted to use different values for the numbers of hidden neurons and inputs and then presented the performance results. The developed model can be used in future studies to be done on fuel consumption and energy efficiency for ships in maritime transport.

Sensor BMP280 Statistical Analysis for Barometric Pressure Acquisition

BMP280 sensor is widely used to measure air pressure. Accuracy and precision are parameters used to indicate the performance of a sensor. This study examined the BMP280 sensor performance by measuring the air pressure on 15 sensors compared with BMKG devices as a reference standard. This measurement was conducted in BMKG class III Tanjungpinang from 20-22 May 2022. BMP280 Sensor performance analysis was performed using a statistical analysis one-way ANOVA method, Tukey test, linear regression, and root mean square error. Based on the results of one-way ANOVA, it is found that there is a BMP280 sensor that is different from the others. The Tukey test found that ID sensors 6 & 14 are not significantly different, ID sensors 1 & 4 are not significantly different, sensors 15 & BMKG are not significantly different, and sensors 3 and 9 are not significantly different. In addition, the results of linear regression and RMSE show an increase in the accuracy of the BMP280 sensor reading. Linear regression makes sensor measurement more accurate due to the small RMSE value.

Variational iteration method along with intelligent computing system for the radiated flow of electrically conductive viscous fluid through porous medium

This article aims to investigate the analytical nature and approximate solution of the radiated flow of electrically conductive viscous fluid into a porous medium with slip effects (RFECVF). In order to build acceptable accurate solutions for RFECVF, this study presented an efficient Levenberg-Marquardt technique of artificial neural networks (LMT-ANNs) approach. One of its fastest back-propagation algorithms for nonlinear lowest latency is the LMT. To turn a quasi-network of PDEs expressing RFECVF into a set of standards, the appropriate adjustments are required. During the flow, the boundary is assumed to be convective. The flow and heat transfer are governed by partial differential equations, and similarity transform is the main tool to convert it into a coupled nonlinear system of ODEs. The usefulness of the constructed LMT-ANNs for such a modelled issue is demonstrated by the best promising algebraic outputs in the E−03 to E−08 range, as well as error histogram and regression analysis measures. Mu is a controller that oversees the entire training procedure. The LMT-ANNs mainly focuses on the higher accuracy of nonlinear systems. Analytical results for the improved boundary layer ODEs are produced using the Variational Iteration Method, a tried-and-true method (VIM). The Lagrange Multiplier is a powerful tool in the suggested method for reducing the amount of computing required. Further, a tabular comparison is provided to demonstrate the usefulness of this study. The final results of the Variational Iteration Method (VIM) in MATLAB have accurately depicted the physical characteristics of a number of parameters, including Eckert, Prandtl, Magnetic, and Thermal radiation parameters.

Regularized fingerprinting in detection and attribution of climate change with weight matrix optimizing the efficiency in scaling factor estimation

Detection and attribution analyses play a central role in establishing the causal effect of human activities on global warming. The most commonly used method in such analyses, optimal fingerprinting, is a multiple regression where each covariate has a measurement error whose covariance matrix is the same as that of the regression error up to a known scale. Inferences about the regression coefficients are critical not only for making statements about detection and attribution but also for quantifying the uncertainty in important outcomes derived from detection and attribution analyses. When there are no errors-in-variables (EIV), the optimal weight matrix in estimating the regression coefficients is the precision matrix of the regression error. This matrix, however, is never known and has to be estimated from climate model simulations with regularization. The consequence is that the optimal fingerprinting method is not optimal, as believed in practice. We construct a weight matrix by inverting a nonlinear shrinkage estimate of the error covariance matrix that minimizes loss functions directly targeting the uncertainty of the resulting regression coefficient estimator. The resulting estimator of the regression coefficients is asymptotically optimal as the sample size of the climate model simulations and the matrix dimension go to infinity together with a limiting ratio. When EIVs are present, the estimator of the regression coefficients, based on the proposed weight matrix, is asymptotically more efficient than that based on the inverse of the existing linear shrinkage estimator of the error covariance matrix. The performance of the method is confirmed in finite sample simulation studies mimicking realistic situations in terms of the length of the confidence intervals and empirical coverage rates for the regression coefficients. In an application to detection and attribution analyses of the mean temperature at different spatial scales, the method yielded shorter confidence intervals which are important for such analyses in practice.

Translation-invariant functional clustering on COVID-19 deaths adjusted on population risk factors

Abstract This paper focuses on clustering the COVID-19 death rates reported in Europe and the United States. Several methods have been developed to cluster such functional data. However, these methods are not translation-invariant (TI) and thus cannot handle different times of arrivals of the disease, nor can they consider external covariates and so are unable to adjust for the population risk factors of each region. We propose a novel three steps clustering method to circumvent these issues. First, feature extraction is performed by TI wavelet decomposition, which permits to deal with the different onsets. Then, single-index regression is used to neutralize disparities caused by population risk factors. Finally, a nonparametric mixture is fitted on the regression residuals to achieve the region clustering.

Generalized residuals and outlier detection for ordinal data with challenging data structures

Motivated by the analysis of rating data concerning perceived health status, a crucial variable in biomedical, economic and life insurance models, the paper deals with diagnostic procedures for identifying anomalous and/or influential observations in ordinal response models with challenging data structures. Deviations due to some respondents’ atypical behavior, outlying covariates and gross errors may affect the reliability of likelihood based inference, especially when non robust link functions are adopted. The present paper investigates and exploits the properties of the generalized residuals. They appear in the estimating equations of the regression coefficients and hold the remarkable characteristic of interacting with the covariates in the same fashion as the linear regression residuals. Identification of statistical units incoherent with the model can be achieved by the analysis of the residuals produced by maximum likelihood or robust M-estimation, while the inspection of the weights generated by M-estimation allows to identify influential data. Simple guidelines are proposed to this end, which disclose information on the data structure. The purpose is twofold: recognizing statistical units that deserve specific attention for their peculiar features, and being aware of the sensitivity of the fitted model to small changes in the sample. In the analysis of the self-perceived health status, extreme design points associated with incoherent responses produce highly influential observations. The diagnostic procedures identify the outliers and assess their influence.

Digital Inclusion, Cultural Capital, and Health Status of Urban and Rural Residents: An Empirical Study Based on 2017 CGSS Database.

China is committed to using digital technology to drive urban-rural integration in health care. This study aims to explore the effect of digital inclusion on health status with the mediating role of cultural capital and the digital health disparities between urban and rural residents in China. Using data from the 2017 Chinese General Social Survey (CGSS), the present study adopted an ordinary least squares (OLS) robust standard error regression model to investigate the impact of digital inclusion on health status. In addition, causal step regression (CSR) and bootstrapping methods were combined to test the mediating effect of cultural capital. The results showed that, first, digital inclusion was related to positive and significant effects on resident health status. Second, cultural capital played a mediating role in the relationship between digital inclusion and health status. Third, urban residents gained more health benefits from digital inclusion than rural residents. Additionally, common method variance (CMV) tests, endogenous tests, and a propensity score matching (PSM) analysis showed that the above conclusions remained robust. The government should therefore focus not only on promoting the population's health by utilizing digital inclusion but also on accelerating digital health equity between urban and rural areas by developing such strategies as a digital infrastructure expansion schedule and digital literacy education and training programs.

Optimization of Chaboche Material Parameters with a Genetic Algorithm

The main objective of this study is to research and develop a genetic algorithm (GA) for optimizing Chaboche material model parameters within an industrial environment. The optimization is based on 12 experiments (tensile, low-cycle fatigue, and creep) that are performed on the material, and corresponding finite element models were created using Abaqus. Comparing experimental and simulation data is the objective function that the GA is minimizing. The GA's fitness function makes use of a similarity measure algorithm to compare the results. Chromosome genes are represented with real-valued numbers within defined limits. The performance of the developed GA was evaluated using different population sizes, mutation probabilities, and crossover operators. The results show that the population size had the most significant impact on the performance of the GA. With a population size of 150, a mutation probability of 0.1, and two-point crossover, the GA was able to find a suitable global minimum. Comparing it to the classic trial and error approach, the GA improves the fitness score by 40%. It can deliver better results in a shorter time and offer a high degree of automation not present in the trial and error approach. Additionally, the algorithm is implemented in Python to minimize the overall cost and ensure its upgradability in the future.

Developing explicit customer preference models using fuzzy regression with nonlinear structure

In online sales platforms, product design attributes influence consumer preferences, and consumer preferences also have a significant impact on future product design optimization and iteration. Online review data are the most intuitive feedback from consumers on products. Using the value of online review information to explore consumer preferences is the key to optimize the products, improve consumer satisfaction and meet consumer requirements. Therefore, the study of consumer preferences based on online reviews is of great importance. However, in previous research on consumer preferences based on online reviews, few studies have modeled consumer preferences. The models often suffer from the nonlinear structure and the fuzzy coefficients, making it challenging to build explicit models. Therefore, this study adopts a fuzzy regression approach with a nonlinear structure to model consumer preferences based on online reviews to provide reference and insight for subsequent studies. First, smartwatches were selected as the research object, and the sentiment scores of product reviews under different topics were obtained by text mining on the product online data. Second, a polynomial structure between product attributes and consumer preferences was generated to investigate the association between them further. Afterward, based on the existing polynomial structure, the fuzzy coefficients of each item in the structure were determined by the fuzzy regression approach. Finally, the mean relative error and mean systematic confidence of the fuzzy regression with nonlinear structure method were numerically calculated and compared with fuzzy least squares regression, fuzzy regression, adaptive neuro fuzzy inference system (ANFIS) and K-means-based ANFIS, and it was found that the proposed method was relatively more effective in modeling consumer preferences.