Semiparametric Regression Approach Research Articles

MODELLING OF POVERTY PERCENTAGE IN EAST JAVA PROVINCE WITH SEMIPARAMETRIC REGRESSION APPROACH

Poverty is an economic problem faced by all countries in the world, including Indonesia. Poverty is seen as the inability of a person from an economic standpoint to meet basic food and non-food needs as measured from the expenditure side. East Java Province is used as the object of research because this province has the highest economic growth in Java Island after DKI Jakarta province in the last 5 years. However, East Java is also included in the province with the highest number of poor people on the island of Java. Several independent variables that are thought to influence the percentage of poverty in East Java are the Open Unemployment Rate (TPT), Life Expectancy Rate (AHH), Average Years of Schooling (RLS), Population Density, and GRDP Rate. Sources of research data come from the East Java BPS website and East Java Open Data. Data analysis was performed using a semiparametric regression approach. The results of the analysis obtained good performance values, namely the MSE value of 12,2156 and the R2 value of 98,71%.

From little girls to adult women: Changes in age at marriage in Scheduled Castes from Madhya Pradesh and Uttar Pradesh, India.

Research confirms the negative relationship between early marriage and mothers' and children's health outcomes. This is why studies of the changes in age at marriage are an important task from the point of view of the health status and well-being of a mother and her offspring, especially in groups represented by extremely disadvantaged social strata in India. The results of such studies may influence the future family planning policy in the country. This study aims to investigate the trend of age at marriage among the Scheduled Castes (SCs) women from two Indian states: Madhya Pradesh and Uttar Pradesh relative to the level of education and also to socioeconomic changes in the states. These states manifest the highest proportion of girls getting married below the age of 18 years-far above the proportion observed in entire India. Women from Scheduled Caste, N = 1,612, aged 25-65, born in 1950-1990 were investigated. A modern semiparametric regression approach was used. To capture the relationship between age at marriage and year of birth, categories of women's level of education (illiterate; primary: 1st-5th standards; middle school: 6th-8th standards; high school: 9th-10th standards; higher secondary: 11th-12th standards), and categories of the profession (women working in the agricultural sector or the non-agricultural sector), flexible framework of the Generalized Additive Model (GAM) was applied. A significant impact of the cohort defined by the year of birth (<0.001), and women's education (<0.001) on age at marriage was noted, while the influence of women's occupation was not significant (p = 0.642). Mean age at marriage differed significantly with different education level. Women who graduated from primary school married 0.631 years later on average than illiterate ones, while those who graduated from middle schools, high schools (9th-10th standards) and higher secondary schools married significantly later than illiterate ones by 1.454 years and 2.463 years, respectively. Age at marriage increased over time: from slightly above 15 years in the cohort of illiterate women born in 1950 to almost 19 years in quite well-educated women born in 1990. The average age at marriage estimated for four education levels in 1990 ranged between 16.39 years (95%CI: 15.29-17.50) in the group of illiterate women and 18.86 years (95%CI: 17.76-19.95) in women graduated from high and higher secondary schools. The rise of age at marriage can be partly explained by the increase of females enrolled in schools, the alleviation of poverty, and the implementation of social programs for women.

Light as a controlling factor of winter phytoplankton in a monomictic reservoir

Factors affecting the seasonal succession of plankton communities in freshwater temperate lakes have been thoroughly studied for decades. However, there are still relatively few data describing the winter season patterns in detail, as the focus has been mostly on spring to autumn conditions. Ice cover is often the crucial factor limiting light availability for winter phytoplankton, but in warm monomictic lakes is usually lacking and the gradually increasing solar radiation should, theoretically, drive phytoplankton growth. In this study conducted in 2002–2010, we documented regular sharp increases in phytoplankton chlorophyll a, starting just after the winter solstice and lasting throughout the total circulation and/or unstable inverse stratification period in the monomictic Slapy reservoir (Czechia). Chlorophyll a concentrations analysed in one-week intervals reached their yearly minimum of 0.2–0.8 µg L−1 in the solstice period, and the spring peak occurred before the onset of stable thermal stratification. The regular pattern was slightly disrupted in some years, associated with short periods of ice cover. Winter phytoplankton were species poor and dominated by diatoms, cryptophytes, green algae, and cyanobacteria. Using semiparametric regression approach, we aimed to test if selected environmental parameters had a significant effect on the observed winter trend. The resulting model revealed that solar radiation and water temperature positively influenced log chlorophyll a concentrations, whereas water age had a significant negative effect. On the other hand, zooplankton density and ice cover effects were not significant. The shapes of the marginal effects of water temperature and solar radiation were nonlinear, and the interaction of these two major factors was significant. The model-based estimated chlorophyll a concentrations showed a shift from radiation dominance to temperature-positive effects along the temperature gradient. This might represent as yet neglected pattern of phytoplankton seasonal development in warm monomictic lakes worldwide.

A deep learning semiparametric regression for adjusting complex confounding structures

Deep Treatment Learning (deepTL), a robust yet efficient deep learning-based semiparametric regression approach, is proposed to adjust the complex confounding structures in comparative effectiveness analysis of observational data, for example, electronic health record (EHR) data in which complex confounding structures are often embedded. Specifically, we develop a deep learning neural network with a score-based ensembling scheme for flexible function approximation. An improved semiparametric procedure is further developed to enhance the performance of the proposed method under finite sample settings. Comprehensive numerical studies have demonstrated the superior performance of the proposed methods, as compared with existing methods, with a remarkably reduced bias and mean squared error in parameter estimates. The proposed research is motivated by a postsurgery pain study, which is also used to illustrate the practical application of deepTL. Finally, an R package, “deepTL,” is developed to implement the proposed method.

Stochastic prognostics under multiple time-varying environmental factors

Prediction of the remaining useful life of in-field components, traditionally, relies on condition monitoring signals which are correlated with the physical degradation of the system. Many models assume that condition monitoring signals behave under similar environmental conditions (e.g. pressure, temperature, workload and relative humidity) or these conditions have no effect on degradation process. In this paper, we propose a Brownian motion process with a stress-dependent drift to model multiple time-varying environmental covariates. A semiparametric regression approach utilizing penalized splines is, further, proposed to model the environmental covariates-drift relationship. The unique feature of our approach is that it does not assume a functional form for the degradation process drift and models multiple environmental covariates’ effect on the degradation process. Moreover, the model is combined with in situ degradation measurements of the in-field unit and its environmental conditions to predict the unit’s remaining useful life through a Bayesian updating scheme. The performance of the proposed framework is investigated and benchmarked through analysis based on numerical studies and a case study using real-world data of frying oil degradation collected from connected fryers.

Geoadditive Quantile Regression Model for Sewer Pipes Deterioration Using Boosting Optimization Algorithm

Proactive management of wastewater pipes requires the development of deterioration models that support maintenance and inspection prioritization. The complexity and the lack of understanding of the deterioration process make this task difficult. A semiparametric Bayesian geoadditive quantile regression approach is applied to estimate the deterioration of wastewater pipe from a set of covariates that are allowed to affect linearly and nonlinearly the response variable. Categorical covariates only affect linearly the response variable. In addition, geospatial information embedding the unknown and unobserved influential covariates is introduced as a surrogate covariate that capture global autocorrelations and local heterogeneities. Boosting optimization algorithm is formulated for variable selection and parameter estimation in the model. Three geoadditive quantile regression models (5%, 50% and 95%) are developed to evaluate the band of uncertainty in the prediction of the pipes scores. The proposed model is applied to the wastewater system of the city of Calgary. The results show that an optimal selection of covariates coupled with appropriate representation of the dependence between the covariates and the response increases the accuracy in the estimation of the uncertainty band of the response variable. The proposed modeling approach is useful for the prioritization of inspections and provides knowledge for future installations. In addition, decision makers will be informed of the probability of occurrence of extreme deterioration events when the identified causal factors, in the 5% and 95% quantiles, are observed on the field.

Three form fourier series estimator semiparametric regression for longitudinal data

Analysis of regressionis one technique that is often used in statistical analysis. There are three regression analysis approaches, such as parametric regression, nonparametric regression and semiparametric regression. Semiparametric regression consists of parametric components and nonparametric components. Parametric component that used such as linear estimator and nonparametric component by using a Fourier series estimator. Semiparametric regression approach that use Fourier series, have an advantages which is can resolve oscillation data pattern. This study compares the three Fourier series estimators such as sine, cosine, and combination between cosine and sine or complete estimator for longitudinal data. Longitudinal data can explain more complete information than cross section data or time series data. The purpose of this study is to introduce another Fourier series for the application of electricity consumption in Madura island. The results of this study indicated the optimal model in predicting electricity consumption in Madura island. The best estimator is the Fourier series estimator with the smallest Generalized Cross Validation (GCV) and Mean Square Error (MSE), and the biggest determination coefficient values by considering the parsimony of the model.

Long-term trends in fog occurrence in the Czech Republic, Central Europe

Fog is a very important and complex atmospheric phenomenon of the utmost importance for the environment and for human society. For practical reasons, fog occurrence is observed regularly at meteorological stations worldwide. Decreasing trends in fog frequency reported from numerous regions have been often associated with either decreasing pollution or climate change, including increasing temperature and changes in atmospheric circulation.We have examined the data on fog occurrence from twelve Czech sites representing different environments (urban, rural, mountain), geographical areas, and altitudes across the country. For our analysis we used long-term records from the time period of 1961–2018, covering both the ambient air’s heavily polluted periods of the 1970s and 1980s and the cleaner period, following the adoption of new, more stringent legislation and effective countermeasures after the 1990s. We applied a generalised additive model (GAM) framework as a flexible, semiparametric regression approach to address nonlinear trend shapes in a formalised and unified way. In particular, we employed a penalised spline approach with cross-validated penalty coefficient estimation.Our study confirmed non-linear behaviour for both year-to-year trends and annual seasonality. Our results showed further that over the analysed, almost sixty-year period, fog occurrence has decreased significantly at all the examined sites, though the pattern of the long-term change differed among individual sites. Moreover, we have found significant seasonality in fog occurrence, though it is different at individual sites. Furthermore, apart from the overall annual fog probability change over the years, at some sites the fog’s seasonal profile has also deformed substantially over the long term.

Parameter Interval Estimation of Semiparametric Spline Truncated Regression Model for Longitudinal Data

Regression analysis is one method in statistics that is used to determine the pattern of functional relationships between response variables with predictor variables. Semiparametric regression approach is a combination of parametric regression and nonparametric regression. The most popular estimator for nonparametric regression or semiparametric regression is spline truncated estimator. Spline is the estimation method that is most often used because it has excellent statistical interpretation and visual interpretation compared to other methods. Regression modelling using longitudinal data is often found in everyday life, where observations are carried out for each subject over a period of time. Interval estimation is often examined by nonparametric regression and semiparametric regression; this estimation aims to determine predictor variables that have a significant influence on the response variable. One indicator used in poverty analysis is the poverty line. Based on Indonesia’s macro poverty analysis calculations, in the period March 2016 to March 2017, the poverty line increased by 5.67 percent, with increases in urban and rural areas at 5.79 percent and 5.19 percent respectively. Modelling using semiparametric spline truncated regression for longitudinal data on data on the percentage of poor people in Indonesia produces the best model using W1 weighting and one point knot. Based on the results of the study with a significance level of 0.05, it was found that the percentage of poor people was influenced by the human development index (HDI) and the unemployment rate. This semiparametric regression model has a minimum GCV value of 1.677, MSE of 5.477 × 10−2 and R2 value of 98.67%.

Real estate VaR estimation in Seoul and Busan, Korea

In this study, we estimtate the value-at-risk (VaR) of the 10-day average apartment prices of Gangnam-gu, Seoul and Haeundae-gu, Busan in Korea. For this purpose, we adopt the semiparametric quantile regression approach for the VaR calculation based on ARIMA-GARCH models, employing the well-known risk measures such as the conditional autoregressive value-at-risk (CAViaR) and conditional autoregressive expectile (CARE) methods. After conducting the unconditional coverage (UC) and conditional coverage (CC) tests on the estimated VaRs, we conclude that the two methods perform similarly for the Seoul case but the CARE method shows more stability than the CAViaR method in the Busan case.

Quantile forecasts for financial volatilities based on parametric and asymmetric models

For financial volatilities such as realized volatility and volatility index, a new parametric quantile forecast strategy is proposed, focusing on forecast interval and value at risk (VaR) forecast. This fully addresses asymmetries in 3 parts: mean, volatility and distribution. The asymmetries are addressed by the LHAR (leverage heterogeneous autoregressive) model of McAleer and Medeiros (2008) and Corsi and Reno (2009) for the mean part, by the EGARCH model for the volatility part, and by the skew-t distribution for the error distribution part. The method is applied to the realized volatilities and the volatility indexes of the US S&P 500 index, the US NASDAQ index, the Korea KOSPI index in which significant asymmetries are identified. Considerable out-of-sample forecast improvements of the forecast interval and VaR forecast are demonstrated for the volatilities: forecast intervals of volatilities have better coverages with shorter lengths and VaR forecasts of volatility indexes have better violations if asymmetries are properly addressed rather than ignored. The proposed parametric method reveals considerably better out-of-sample performance than the recently proposed semiparametric quantile regression approach of Zikes and Barunik (2016).

Changing temperature extremes based on CMIP5 output via semi‐parametric quantile regression approach

The Expert Team on Climate Change Detection and Indices (ETCCDI) proposed a set of 27 core indices, describing extreme weather and climate events based on daily temperature and precipitation data. There are six percentile‐based temperature extreme indices: four exceedance rates (ERs) (TN10p, TN90p, TX10p and TX90p) and two durations (cold spell duration [CSDI] and warm spell duration [WSDI]), derived by using percentiles for calendar days during the base period 1961–1990 as thresholds. The widely used empirical quantile (percentile) estimator or its bootstrap resampling adjustment may result in inhomogeneity in the derived annual ER series or in their linear trends, as well as seasonally varying biases in the monthly ER series. We present a new data set of the six indices for the historical and three representative concentration pathways (RCP2.6, RCP4.5 and RCP8.5) scenarios simulated by 19 global climate models participating the Coupled Model Intercomparison Project Phase 5, generated via the semi‐parametric quantile regression approach. Percentiles estimated upon the independence condition, inhomogeneity and seasonal biases are removed from the new data set, so that the changes in temperature extremes can be authentically revealed from the climatological distribution point of view. Compared with its counterpart produced by the CLIMDEX project, the new data set shows similar spatial and temporal change patterns under the RCP scenarios, but with much smaller magnitudes. By the end of 21st century, TN90p increases to 17.3, 24.5 and 40.0%, and TX90p increases to 16.9, 22.1 and 35.3%, under RCP2.6, RCP4.5 and RCP8.5, respectively. WSDI increases to 35 days under RCP8.5. For the three increasing indices, the most significant changes occur in tropical and extratropical regions. The greatest increases in TN90p and TX90p occur in southeast Asia and Amazon Basin, up to 85 and 68% in the summer under RCP8.5; that in WSDI occurs in Amazon Basin up to 120 days under RCP8.5.

Age and sex differences in intra-individual variability in a simple reaction time task

While age effects in reaction time (RT) tasks across the lifespan are well established for level of performance, analogous findings have started appearing also for indicators of intra-individual variability (IIV). Children are not only slower, but also display more variability than younger adults in RT. Yet, little is known about potential moderating sex effects on RT-IIV. We analyzed responses in a simple RT task with 120 trials in children, younger, and older adults. To best capture sex differences we used generalized additive models (GAMs), a semi-parametric regression approach, to fit splines relating nonlinearly age to RT, and capable of testing sex differences therein. This method is more adequate to test sex differences in nonlinear age relations than polynomial regression. Results show that (a) males are faster than females (except in the older adults), and (b) in younger and older adults, males are less variable than females. No sex difference in IIV emerged in children. Results are consistent with the hypothesis that sex differences in RT variability may be attributable to brain effects of sex hormones, in particular estrogen, whose receptors are present in several brain regions involved in information processing and attention, which are systems involved in the regulation of variability in information processing. Thus, according to this hypothesis, sex differences in RT-IIV should be present after puberty, but not in pre-pubertal children.

A New Function for Generating Local Bandwidths for Semi-Parametric MRR2 Model in Response Surface Methodology

The modeling phase of response surface methodology (RSM) involves the application of regression techniques to fitting a curve to the data generated from a designed experiment. The model-robust regression 2 (MRR2) method is a semi-parametric regression approach that combines portions of estimates from both the parametric and the nonparametric regression approaches via a mixing parameter. However, the robustness of the estimates from the MRR2 approach depends largely on the choice of bandwidth. Utilizing the cross-validation approach to bandwidth selection, we propose a methodology for deriving a data-driven function that generates local bandwidths for the MRR2 approach. Using two examples from the literature and a simulation study, we show that, in comparison with other regression methods, the results obtained from the MRR2 approach utilizing the proposed function offer remarkable improvements in the goodness-of-fit statistics.

Bayesian Semiparametric Model for Pathway-Based Analysis with Zero-Inflated Clinical Outcomes

In this paper, we propose a semiparametric regression approach for identifying pathways related to zero-inflated clinical outcomes, where a pathway is a gene set derived from prior biological knowledge. Our approach is developed by using a Bayesian hierarchical framework. We model the pathway effect nonparametrically into a zero-inflated Poisson hierarchical regression model with an unknown link function. Nonparametric pathway effect was estimated via a kernel machine, and the unknown link function was estimated by transforming a mixture of the beta cumulative density function. Our approach provides flexible nonparametric settings to describe the complicated association between gene expressions and zero-inflated clinical outcomes. The Metropolis-within-Gibbs sampling algorithm and Bayes factor were adopted to make statistical inferences. Our simulation results support that our semiparametric approach is more accurate and flexible than zero-inflated Poisson regression with the canonical link function, which is especially true when the number of genes is large. The usefulness of our approach is demonstrated through its applications to the Canine data set from Enerson et al. (Toxicol Pathol 34:27–32, 2006). Our approach can also be applied to other settings where a large number of highly correlated predictors are present.

A kernel-based mixed effect regression model for earthquake ground motions

This paper presents a semi-parametric mixed-effect regression approach for analyzing and modeling earthquake ground motions, taking into account the correlations between records. Using kernels, the proposed method extends the classical mixed model equations to complicated relationships. The predictive equation is composed of parametric and nonparametric parts. The parametric part incorporates known relationships into the model, while the nonparametric part captures the relationships which cannot be cast into a simple parametric form. A least squares kernel machine is used to infer the nonparametric part of the model. The resulting semi-parametric model combines the strengths of parametric and nonparametric approaches, allowing incorporation of prior, well-justified knowledge into the model while retaining flexibility with respect to the explanatory variables for which the functional form is uncertain. Equations for pointwise confidence and prediction intervals around the conditional mean are provided. The validity of the proposed method is demonstrated through numerical simulations and using recorded ground motions.

Spatial regression with covariate measurement error: A semiparametric approach.

Spatial data have become increasingly common in epidemiology and public health research thanks to advances in GIS (Geographic Information Systems) technology. In health research, for example, it is common for epidemiologists to incorporate geographically indexed data into their studies. In practice, however, the spatially defined covariates are often measured with error. Naive estimators of regression coefficients are attenuated if measurement error is ignored. Moreover, the classical measurement error theory is inapplicable in the context of spatial modeling because of the presence of spatial correlation among the observations. We propose a semiparametric regression approach to obtain bias-corrected estimates of regression parameters and derive their large sample properties. We evaluate the performance of the proposed method through simulation studies and illustrate using data on Ischemic Heart Disease (IHD). Both simulation and practical application demonstrate that the proposed method can be effective in practice.

Innovation and job creation: a sustainable relation?

This study compared the employment growth patterns of innovative and non-innovative firms, focusing on whether or not there are systematic differences between these two categories in the persistence of the jobs they create. To this end, a unique longitudinal dataset of 3304 Spanish firms over the period 2002–2009 and a semi-parametric quantile regression approach was used. The empirical results indicate that, ceteris paribus, innovative, smaller and younger firms are more likely to experience high employment growth episodes than non-innovative firms. More interestingly, among those firms that contribute more to yearly job creation (e.g. high-growth firms), only innovative companies are able to sustain high growth over time (in contrast to non-innovative firms). In addition, among declining firms, non-innovators tend to deteriorate faster in terms of economic performance than innovators.

Semiparametric Autoregressive Conditional Duration Model: Theory and Practice

Many existing extensions of the Engle and Russell's (1998) Autoregressive Conditional Duration (ACD) model in the literature are aimed at providing additional flexibility either on the dynamics of the conditional duration model or the allowed shape of the hazard function, i.e., its two most essential components. This article introduces an alternative semiparametric regression approach to a nonlinear ACD model; the use of a semiparametric functional form on the dynamics of the duration process suggests the model being called the Semiparametric ACD (SEMI–ACD) model. Unlike existing alternatives, the SEMI–ACD model allows simultaneous generalizations on both of the above-mentioned components of the ACD framework. To estimate the model, we establish an alternative use of the existing Bühlmann and McNeil's (2002) iterative estimation algorithm in the semiparametric setting and provide the mathematical proof of its statistical consistency in our context. Furthermore, we investigate the asymptotic properties of the semiparametric estimators employed in order to ensure the statistical rigor of the SEMI–ACD estimation procedure. These asymptotic results are presented in conjunction with simulated examples, which provide an empirical evidence of the SEMI–ACD model's robust finite-sample performance. Finally, we apply the proposed model to study price duration process in the foreign exchange market to illustrate its usefulness in practice.

Local Bandwidths for Improving Performance Statistics of Model-Robust Regression 2

Model-Robust Regression 2 (MRR2) method is a semi-parametric regression approach that combines parametric and nonparametric fits. The bandwidth controls the smoothness of the nonparametric portion. We present a methodology for deriving data-driven local bandwidth that enhances the performance of MRR2 method for fitting curves to data generated from designed experiments.