Nonparametric Regression Approach Research Articles

Stunting Prevalence Modeling Using Nonparametric Regression of Quadratic Splines

The nonparametric regression approach is used when the shape of the regression curve between the response variable and the predictor variable is assumed to be of unknown shape. The advantages of nonparametric regression have high flexibility. A nonparametric regression approach that is often used is truncated spline which has an excellent ability to handle data whose behavior changes at certain sub-sub intervals. The purpose of this study is to obtain the best model of multivariable nonparametric regression with linear and quadratic truncated spline approaches using the Generalized Cross Validation (GCV) and Unbiased Risk (UBR) methods and to find out the factors that influence the prevalence of stunting in Indonesia in 2021. The data used were the prevalence of stunting as a response variable and the predictor variable used was the percentage of infants receiving exclusive breastfeeding for 6 months, the percentage of households that have proper sanitation, the percentage of toddlers who get Early Initiation of Breastfeeding (IMD), the percentage of poor people, and the percentage of pregnant women at risk of SEZ. The results showed that the best quadratic truncated spline nonparametric regression model in modeling stunting prevalence was quadraic truncated spline using the GCV method with three knot points. This model has a minimum GCV value of 7.29 with an MSE value of 1.82 and a R2 value of 94.07%.

MODELING CHRONIC FILARIASIS CASES IN WEST JAVA USING A MULTIVARIATE ADAPTIVE APPROACH REGRESSION SPLINES

One of the most crippling infectious diseases in the world is filariasis. Indonesia is a unitary country with 34 provinces, where West Java is one of the 5 provinces with the most filariasis sufferers in Indonesia as of 2021. Reinfection occurs in places that have implemented POMP. Therefore, monitoring operations must be carried out to track the emergence of new cases and risk factors for transmission. The aim of this research focuses on describing and modeling the number of chronic filariasis in West Java, as well as interpreting the best model results obtained. The method used is a method with a nonparametric regression approach, namely Multivariate Adaptive Regression Spline. The results of the research show that the best model obtained is a combination of 15 base functions, maximum interaction 2, and minimum observation between knots 1. From this model, the predictor variable that has the most influence on the response variable in order based on the level of variable importance is the Percentage of Population Access to Facilities Decent Sanitation, Percentage of Households with Clean and Healthy Behavior (PHBS), Sex Ratio, and Percentage of Poor Population. The interpretation of the best model is that the variable Percentage of Population Access to Adequate Sanitation Facilities above 6,650% will contribute to a reduction in the number of chronic filariasis; the Sex Ratio variable below 103,300 will contribute in the form of a reduction in the number of chronic filariasis. it can be seen that the predictor variable that has the most influence on the response variable is the variable Percentage of Population Access to Proper Sanitation Facilities with an importance level of 100%.

MODELLING CRIME RATES IN INDONESIA USING TRUNCATED SPLINE ESTIMATOR

Criminal acts are actions that violate the law and can arise from various factors such as emotions, psychological pressure, and others. Crime rate is a number that indicate the level of crime vulnerability in a certain area at a certain time. Higher crime rates correspond to increased vulnerability in an area, and vice versa. Among various forms of criminal acts, the number of criminal acts and narcotics crimes in Indonesia tends to increase in 2020 and 2021. The aim of the research is to identify the characteristics of crime rate data based on the number of decency and narcotics incidents in Indonesia using a nonparametric regression approach. This research uses a nonparametric regression method spline truncated and linear regression as comparison. It was found that West Papua Province has the highest crime rate, based on a comparison between linear regression model and truncated spline nonparametric regression model, it can be concluded that the best model is the truncated spline nonparametric regression model with a Generalized Cross Validation (GCV) of 2468.487 and a coefficient of determination of 0 .7389091, indicating that approximately 73% of the variability of the dependent variable can be explained by the independent variables included in the model.

APPLICATION OF NONPARAMETRIC REGRESSION SPLINE TRUNCATED FOR MODELING THE HEIGHT OF YEOP CHAGI KICKS OF TAEKWONDO ATHLETES IN SAMARINDA CITY

Nonparametric regression is a model approach method that is used when the shape of the regression curve between the response variable and the predictor variable is assumed to have an unknown shape or pattern. One of the estimators in the nonparametric regression approach is the truncated spline which has the ability to handle data whose behavior changes at certain sub intervals. The purpose of this study was to obtain the estimated value of the parameters of the nonparametric regression model with a truncated spline approach at one knot point, two knot points, and three knot points for kick height data of yeop chagi taekwondo athletes in Samarinda City. The results showed that the truncated spline nonparametric regression model was the best in modeling high kick height data for yeop chagi taekwondo athletes in Samarinda City with three knot points. This model has the minimum Generalized Cross Validation (GCV) value of 7.94 with an R2 value of 94.72% and a Mean Square Error (MSE) value of 2.62. Based on the results of the model parameter significance test, it was concluded that the factors that influence the kick height of the yeop chagi taekwondo athlete in Samarinda City are flexibility, leg power, leg length, and waist circumference.

Novel Approach to Design a Model for Software Effort Estimation Using Linear Regression

Estimation of an effort is necessary to decide the time and cost needed for the completion of the project. Making judgments with a high degree of uncertainty is a serious issue in the field of software engineering. Software quality forecasting calls for highly precise technologies and expert knowledge. The forecast of software effort based on past data from software development metrics could, instead, be aided by AI-based predictive models, which have a high degree of accuracy. Using a linear regression model, we developed a software effort estimation model in this study to forecast this effort. A non-parametric linear regression approach based on K-Nearest Neighbors was used to create this statistical model (KNN). So, with a 76% coefficient of determination, our findings indicate the potential for applying AI algorithms to predict the software engineering work prediction problem.

Estimation of Multiresponse Multipredictor Nonparametric Regression Model Using Mixed Estimator

In data analysis using a nonparametric regression approach, we are often faced with the problem of analyzing a set of data that has mixed patterns, namely, some of the data have a certain pattern and the rest of the data have a different pattern. To handle this kind of datum, we propose the use of a mixed estimator. In this study, we theoretically discuss a developed estimation method for a nonparametric regression model with two or more response variables and predictor variables, and there is a correlation between the response variables using a mixed estimator. The model is called the multiresponse multipredictor nonparametric regression (MMNR) model. The mixed estimator used for estimating the MMNR model is a mixed estimator of smoothing spline and Fourier series that is suitable for analyzing data with patterns that partly change at certain subintervals, and some others that follow a recurring pattern in a certain trend. Since in the MMNR model there is a correlation between responses, a symmetric weight matrix is involved in the estimation process of the MMNR model. To estimate the MMNR model, we apply the reproducing kernel Hilbert space (RKHS) method to penalized weighted least square (PWLS) optimization for estimating the regression function of the MMNR model, which consists of a smoothing spline component and a Fourier series component. A simulation study to show the performance of proposed method is also given. The obtained results are estimations of the smoothing spline component, Fourier series component, MMNR model, weight matrix, and consistency of estimated regression function. In conclusion, the estimation of the MMNR model using the mixed estimator is a combination of smoothing spline component and Fourier series component estimators. It depends on smoothing and oscillation parameters, and it has linear in observation and consistent properties.

Augmented inverse probability weighted estimation and prediction for cause-specific proportional hazards regression with missing covariates

This paper describes estimation of the regression parameters and prediction of the cumulative incidence functions under the cause-specific proportional hazards model when some of covariates are not fully observed. Assuming that missingness mechanism is missing at random, we propose the augmented inverse probability weighted method for estimation and inference procedures. A nonparametric regression approach is adapted for estimating selection probabilities and conditional expectations of missing covariates in the augmented estimating function. We establish the asymptotic properties of the predicted cumulative incidence functions under the cause-specific proportional hazards model with missing covariates and derive consistent variance estimators of the predicted cumulative incidence functions. Simulation studies show that the procedures perform well. The proposed methods are illustrated with stage IV breast cancer data obtained from the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute.

Dynamic Return Scenario Generation Approach for Large-Scale Portfolio Optimisation Framework

AbstractIn this paper, we propose a complex return scenario generation process that can be incorporated into portfolio selection problems. In particular, we assume that returns follow the ARMA–GARCH model with stable-distributed and skewed t-copula dependent residuals. Since the portfolio selection problem is large-scale, we apply the multifactor model with a parametric regression and a nonparametric regression approaches to reduce the complexity of the problem. To do this, the recently proposed trend-dependent correlation matrix is used to obtain the main factors of the asset dependency structure by applying principal component analysis (PCA). However, when a few main factors are assumed, the obtained residuals of the returns still explain a non-negligible part of the portfolio variability. Therefore, we propose the application of a novel approach involving a second PCA to the Pearson correlation to obtain additional factors of residual components leading to the refinement of the final prediction. Future return scenarios are predicted using Monte Carlo simulations. Finally, the impact of the proposed approaches on the portfolio selection problem is evaluated in an empirical analysis of the application of a classical mean–variance model to a dynamic dataset of stock returns from the US market. The results show that the proposed scenario generation approach with nonparametric regression outperforms the traditional approach for out-of-sample portfolios.

PEMODELAN REGRESI NONPARAMETRIK DENGAN ESTIMATOR SPLINE POLYNOMIAL TRUNCATED PADA DATA JUMLAH WISATAWAN NUSANTARA

The nonparametric regression approach is a statistical method used to determine the relationship between predictor variables and the dependent variable when the assumed pattern is unknown. Truncated spline is an estimator used in nonparametric regression to handle data with varying behaviors. Nonparametric regression modeling with truncated polynomial spline was applied to local Indonesian tourist visitation data obtained from BPS for the years 2017-2019, for each month. The optimal knot points were selected based on the smallest Gross Cross Validation values. Based on the analysis, the optimal model is a second-order spline with the smallest Gross Cross Validation value of 17,95 and the optimal knot points are in the 2nd, 6th, and 7th months. The goodness of the model is evident from an value of 81,88% and an MSE of 12,46. The best model obtained shows a fairly accurate ability to explain the estimated number of domestic tourists so that it can be a basis for stakeholders to make key decisions in planning and managing the tourism industry as an effort to increase domestic tourism interest.

Comparison of Kernel Smoothing and Local Polynomial Smoothing Method in Overcoming Age Heaping

Age data plays an important role in every aspect yet there are found age misreporting. It involves digit preference that causes build up in a certain age. Digit preference in demography is called age heaping that often happens at age with 0 and 5 as the last digit. Age heaping induces poor data quality and data bias that could influence government policy making. Two indicators used to detect age heaping are Whipple Index (WI) and Myers Blended Index (MBI). Methods to cope with age heaping are nonparametric regression approaches which are Kernel Smoothing and Local Polynomial Smoothing. The objective of this research is to measure and elevate the quality of population age data and population mortality data in Sensus Penduduk (SP) 2020 as well as comparing methods between Kernel Smoothing and Local Polynomial Smoothing. The data being used in this paper is SP2020 which the research variables are age population, age of death, and total population. The result shows that the data quality of total population death is inaccurate compared to total population thus needs a smoothing process to improve age data to population data accuration. The method that has better accuracy is the Local Polynomial Smoothing method.

Modeling The Impact Analysis of The Covid-19 Pandemic On The Tourism Sector In Palopo City With A Nonparametric Regression Approach

The implementation of social distancing causes a huge effect on the life sector in Indonesia, one of which is the tourism sector which has experienced the greatest impact due to this policy. One of the treatments that can be done is to find out the factors that are thought to have an influence on the increase in the tourism sector in Palopo city, many methods can be done, one of which is the regression analysis method. This study uses Spline non-parametric regression in modeling the data. The non-parametric regression model was chosen based on results of the identification that the data does not follow a certain distribution pattern. Gross regional domestic product data is used in the model. This can be seen from the resulting value of 76.94%. Gross Regional Domestic Product data is also modeled using multiple linear regression. This modeling is intended as a comparison model. The results show that the use of the non-parametric Spline better model in modeling the data compared to multiple linear regression.

TUBERCULOSIS CASE MODEL USING GCV AND UBR KNOT SELECTION METHODS IN TRUNCATED SPLINE NONPARAMETRIC REGRESSION

The nonparametric regression approach is used when the shape of the regression curve is not known. The advantage of nonparametric regression is that it has a high degree of flexibility. The truncated spline is a method in the nonparametric regression approach, which can overcome changing data patterns at certain sub-intervals with the help of knot points. The purpose of this research is to obtain the best truncated spline nonparametric regression model estimates based on the GCV and UBR knot point selection methodsThe data used in this study came from the publications of the Indonesian Ministry of Health and BPS Indonesia. The response variable used is the percentage of successful treatment of tuberculosis patients in Indonesia with predictor variables namely the percentage of people who smoke over the age of 15 years, the percentage of households that have access to proper sanitation, the percentage of poor people, the percentage of food processing establishments that meet the standard requirements , national health insurance membership coverage and percentage of accredited hospitals. The results showed that the best model came from the GCV method using three knots. This model produces an MSE value of 3.65 with value of 97.04. The value indicates that the predictor variable used in this study affects the response variable by 97.04% while the other 2.96% is influenced by other variables that are not included in this study.

JCI MODELING IN INDONESIA BASED ON INDUSTRIAL PRODUCTION INDEX WITH LOCAL POLYNOMIAL ESTIMATOR APPROACH

The industrial sector is the leading sector that contributes the most to Indonesia's economic growth. Industry can be caused by various factors, one of which is the Jakarta Composite Index (JCI). Indonesian stock prices have a high variance that requires proper modeling. Therefore, this study uses a local polynomial nonparametric regression approach. This study aims to estimate and obtain the best JCI model based on the production index of large and medium industries using a local polynomial estimator and also knowing the accuracy of the JCI model based on the production index of large and medium industries. The data used in this study is secondary data using production index data for medium-large industries and data on the composite stock index in Indonesia in the form of Time series which were obtained through the Central Statistics Agency Publication website on the page www.bps.go.id. JCI modeling in Indonesia based on the production index of large and medium industries is most effective on local polynomials with polynomial degree two which obtains an optimal bandwidth of 7,8795 with a minimum Cross-Validation (CV) value of 163170,3 and a Mean Absolute Percentage Error (MAPE) value of 9,1%. From the MAPE value it is said that the model is good for making future predictions.

Fast and Smooth Composite Local Learning-Based Adaptive Control.

Model structure representation and fast estimation of perturbations are two key research aspects in adaptive control. This work proposes a composite local learning adaptive control framework, which possesses fast and flexible approximation to system uncertainties and meanwhile smoothens control inputs. Local learning, which is a nonparametric regression approach, is able to automatically adjust the structure of approximator based on data distribution from the local region, but it is sensitive to the outliers and measurement noises. To tackle this problem, the regression filter technique is employed to attenuate the adverse effect of noises by smoothing the output response and state features. In addition, the stable integral adaptation is integrated into local learning framework to further enhance the system robustness and smoothness of the estimation. Through the online elimination of uncertainties, the nominal control performance is recovered when the plant encounters violent perturbations. Stability analysis and numerical simulations are performed to demonstrate the effectiveness and benefits of the proposed control method. The proposed approach exhibits a promising performance in terms of rapid perturbation elimination and accurate tracking control.

Kernel Nonparametric Regression for Forecasting Local Original Income

Regional Original Revenue (ROR) is an income collected based on regional regulations under statutory regulations. ROR aims to give authority to Regional Governments to sponsor the implementation of regional autonomy following regional potential. Every year, the Central Lombok Regency government sets ROR targets to assist the government in formulating regional policies. The targets set by the government are sometimes not following their realization. This study aims to determine a model that can be used in forecasting ROR targets. One way to predict the value of ROR is by using a nonparametric regression approach. This approach is flexible since it is not dependent on a particular model. The use of the nonparametric kernel regression method with the Gaussian kernel function obtained a minimum GCV value of 1,769688931 with an optimum bandwidth value of of 0,212740452 and of 0,529682589. Modeling with optimum bandwidth produces a coefficient of determination of 87,55%. The best model is used for forecasting and produces a MAPE value of 5,4%. The analysis results show that what influences the value of ROR is ROR receipts in the previous month and the previous 12 months.

Environmental impact of globalization: The case of central and Eastern European emerging economies

Against the backdrop of piling environmental concerns in the modern era of globalization, this study aims to check the validity of the Pollution Haven Hypothesis (PHH) in Eastern European emerging countries and the relevance of globalization. The study targets to reduce the lack of consensus on the globalization-economic complexity-environment in European countries. Besides, we also intend to explore the existence of an N-shaped economic complexity-related Environmental Kuznets Curve (EKC) controlling for the bearing of renewable energy on environmental degradation. For analytical purposes, both parametric and non-parametric quantile regression approaches are employed. Overall, we find a non-linear relationship between economic complexity and carbon emissions, and N-shaped EKC is verified. Globalization and renewable energy consumption boost and inhibit emissions, respectively. More importantly, the results confirm the moderating role of economic complexity in neutralizing the carbon emissions-boosting effect of globalization.On the other hand, the non-parametric findings show that the N-shaped EKC hypothesis does not hold for high emissions quantiles. Furthermore, for all emissions quantiles, it is found that globalization boosts emissions, economic complexity, and globalization jointly curbs emissions and renewable energy curbs emissions. Based on the overall findings, some vital environmental development policies are recommended. The conclusions support shaping policy options promoting economic complexity and renewable energy as key factors in mitigating carbon emissions.

Stunting determinants among toddlers in Probolinggo district of Indonesia using parametric and nonparametric ordinal logistic regression models

Stunting is a chronic nutritional problem in toddlers characterized by a shorter height than other children of their age. Stunting is a major nutritional problem faced by Indonesia. This research aimed to develop a risk model for the incidence of stunting in toddlers. This research was conducted in the village of stunting locus in the Public Health Center area that was selected to be the sample in Probolinggo District. Data were collected in the villages of Alaspandan, Bucorwetan, Petunjungan, and Sukodadi. The samples taken were 202 toddlers. The results show that the prevalence of a stunted toddler was 26.7% consisting of 21.3% moderately stunting and 5.4% severely stunting, and birth length, maternal height, and health services were important determinants of stunting in toddlers. Also, the value of the classification accuracy of the obtained model using parametric ordinal logistic regression approach was 72.45% which is less than that using nonparametric ordinal logistic regression approach namely 73.98%. It means that the best model for modeling risk of toddlers stunting in Probolinggo District was obtained based on the nonparametric logistic regression approach.

The use of the binary spline logistic regression model on the nutritional status data of children

The nutritional status of children can be grouped into two, namely normal and abnormal nutrition so that the data are analyzed using a binary logistic regression model. In this study, the use of binary logistic regression was developed on the use of the spline estimator as part of the nonparametric regression approach. This model is able to model qualitative response data by showing several trends that may occur in the data. Data on the nutritional status of children were analyzed based on the child's weight through a linear spline estimator and the optimal model was obtained at the use of a one-knot point, which was 4.7 kg. These results indicate that if a child has a body weight of 4.7 kg and above then there is a chance that the child has an abnormal nutritional status of 9.013 times compared to a child with a body weight below 4.7 kg. This could be due to the fact that children are no longer getting breast milk, so they need to get attention regarding their nutritional status.

Modeling of poverty level in South Sulawesi Province through spline nonparametric regression approach

This study discussed modeling poverty levels in South Sulawesi through Spline Nonparametric Regression Approach. Since no specific pattern was formed from the relation between the poverty level in South Sulawesi and the factors that influence it, the researchers used nonparametric regression modelling with Spline approach. The Spline model is very good in modelling data that has changeable patterns at certain subintervals. The aimed of this study were to investigate the factors that have the most influence on the poverty level and modelling the poverty level in South Sulawesi through nonparametric regression. The scope of this study was the use of Generalized Cross-Validation (GCV) in selecting optimal knot points; in this case, it used 1, 2, and 3 knots. This study found that three factors mainly influence the poverty level in South Sulawesi in 2017: unemployment, population growth, and literacy rates. The results showed that the best spline model was the model with three-knot points and the minimum GCV value is 11.1155. In addition, the value is 79.75%. It means that the variables of unemployment, population growth, and literacy rate can explain 79.75% of the variation of the poverty variable, while other variables explain the remaining 20.25%.

Bayesian additive regression trees for multivariate skewed responses.

This paper introduces a nonparametric regression approach for univariate and multivariate skewed responses using Bayesian additive regression trees (BART). Existing BART methods use ensembles of decision trees to model a mean function, and have become popular recently due to their high prediction accuracy and ease of use. The usual assumption of a univariate Gaussian error distribution, however, is restrictive in many biomedical applications. Motivated by an oral health study, we provide a useful extension of BART, the skewBART model, to address this problem. We then extend skewBART to allow for multivariate responses, with information shared across the decision trees associated with different responses within the same subject. The methodology accommodates within-subject association, and allows varying skewness parameters for the varying multivariate responses. We illustrate the benefits of our multivariate skewBART proposal over existing alternatives via simulation studies and application to the oral health dataset with bivariate highly skewed responses. Our methodology is implementable via the R package skewBART, available on GitHub.