Local Regression Method Research Articles

Degradation prognosis for proton exchange membrane fuel cell based on hybrid transfer learning and intercell differences

Proton exchange membrane fuel cell (PEMFC) has been widely used in diverse applications. However, degradation and durability problem is one of the biggest barriers to take PEMFCs into extensive commercial use. Prognostics and health management is an effective solution to this problem. In this study, we focus on its core technology prognostics and propose an individual difference conscious prediction method for PEMFC using a hybrid transfer learning approach to get higher accuracy. Firstly, a time-scale self-optimization local weighted regression method is designed to adaptively smooth the raw data to prominent the performance degradation trend. Then, to obtain a more similar curve to the predicted fuel cell as the training data of the prediction model, a transferability measurement method using cosine-distance selects the most similar historical test data. Furtherly, it is utilized to generate a more similar curve by a data transfer method combining a deep learning model named stacked autoencoder and a hybrid transfer learning strategy. Two types of transfer learning approaches are fused to maximally mine available information from historical data and previous models to help improve the similarity of the generated curve. In this process, the common degradation information of all cells and individual information of the predicted cells are considered to improve generation quality. Finally, a prediction model using stacked Long–short Term Memory(LSTM) having a significant advantage in modeling series relation is trained by the generated samples cut with variable width sliding windows and estimates remaining useful life(RUL) the target fuel cell. Experimental validation data are employed to verify the effectiveness of the proposed algorithm. Satisfying results are also obtained by accuracy comparison under different smoothing scales, numbers of transferable samples, and prediction methods.

Deciphering Sex-Specific Genetic Architectures Using Local Bayesian Regressions.

Many complex human traits exhibit differences between sexes. While numerous factors likely contribute to this phenomenon, growing evidence from genome-wide studies suggest a partial explanation: that males and females from the same population possess differing genetic architectures. Despite this, mapping gene-by-sex (G×S) interactions remains a challenge likely because the magnitude of such an interaction is typically and exceedingly small; traditional genome-wide association techniques may be underpowered to detect such events, due partly to the burden of multiple test correction. Here, we developed a local Bayesian regression (LBR) method to estimate sex-specific SNP marker effects after fully accounting for local linkage-disequilibrium (LD) patterns. This enabled us to infer sex-specific effects and G×S interactions either at the single SNP level, or by aggregating the effects of multiple SNPs to make inferences at the level of small LD-based regions. Using simulations in which there was imperfect LD between SNPs and causal variants, we showed that aggregating sex-specific marker effects with LBR provides improved power and resolution to detect G×S interactions over traditional single-SNP-based tests. When using LBR to analyze traits from the UK Biobank, we detected a relatively large G×S interaction impacting bone mineral density within ABO, and replicated many previously detected large-magnitude G×S interactions impacting waist-to-hip ratio. We also discovered many new G×S interactions impacting such traits as height and body mass index (BMI) within regions of the genome where both male- and female-specific effects explain a small proportion of phenotypic variance (R2 < 1 × 10−4), but are enriched in known expression quantitative trait loci.

Semi-functional partial linear spatial autoregressive model

This paper proposes a semi-functional partial linear spatial autoregressive (SAR) model, in which we allow one of the explanatory variables to be a functional variable, while the dependent variable is scalar. We aim to enable the spatial econometric models to be applied to various areas where data types can be functional data. Based on quasi-maximum likelihood estimation (QMLE) method and local linear regression method, we construct a two-stage estimator to estimate the parameters and nonparametric component. The convergence rate of the estimator of nonparametric component is given. Furthermore, Monte Carlo simulations are performed to investigate our two-stage estimator’s finite sample performance.

Multi-Label Classification Scheme Based on Local Regression for Retinal Vessel Segmentation.

Segmenting small retinal vessels with width less than 2 pixels in fundus images is a challenging task. In this paper, in order to effectively segment the vessels, especially the narrow parts, we propose a local regression scheme to enhance the narrow parts, along with a novel multi-label classification method based on this scheme. We consider five labels for blood vessels and background in particular: the center of big vessels, the edge of big vessels, the center as well as the edge of small vessels, the center of background, and the edge of background. We first determine the multi-label by the local de-regression model according to the vessel pattern from the ground truth images. Then, we train a convolutional neural network (CNN) for multi-label classification. Next, we perform a local regression method to transform the previous multi-label into binary label to better locate small vessels and generate an entire retinal vessel image. Our method is evaluated using two publicly available datasets and compared with several state-of-the-art studies. The experimental results have demonstrated the effectiveness of our method in segmenting retinal vessels.

The European Russia Drought Atlas (1400–2016 CE)

We present the European Russia Drought Atlas (ERDA) that covers the East European Plain to the Ural Mountains from 1400–2016 CE. Like the Old World Drought Atlas (OWDA) for the Euro-Mediterranean region, the ERDA is a one-half degree gridded reconstruction of summer Palmer Drought Severity Indices estimated from a network of annual tree-ring chronologies. Ensemble point-by-point regression is used to generate the ERDA with the identical protocols used for developing the OWDA. Split calibration/validation tests of the ERDA indicate that it has significant skill over most of its domain and is much more skillful than the OWDA where they overlap in the western part of ERDA domain. Comparisons to historical droughts over European Russia additionally support the ERDA’s overall validity. The ERDA has been spatially smoothed and infilled using a local regression method to yield a spatially complete drought atlas back to 1400 CE. EOF analysis indicates that there are three principal modes of hydroclimatic variability in the ERDA. After Varimax rotation, these modes correlate significantly with independent climate data sets extending back to the late nineteenth century in a physically interpretable way and relate to atmospheric circulation dynamics of droughts and heatwaves over European Russia based on more recent instrumental data.

Analyzing Intracellular Gradients in Pollen Tubes.

Conspicuous intracellular gradients manifest and/or drive intracellular polarity in pollen tubes. However, quantifying these gradients raises multiple technical challenges. Here we present a sensible computational protocol to analyze gradients in growing pollen tubes and to filter nonrepresentative time points. As an example, we use imaging data from pollen tubes expressing a genetically encoded ratiometric Ca2+ probe, Yellow CaMeleon 3.6, from which a kymograph is extracted. The tip of the pollen tube is detected with CHUKNORRIS, our previously published methodology, allowing the reconstruction of the intracellular gradient through time. Statistically confounding time points, such as growth arrest where gradients are highly oscillatory, are filtered out and a mean spatial profile is estimated with a local polynomial regression method. Finally, we estimate the gradient slope by the linear portion of the decay in mean fluorescence, offering a quantitative method to detect phenotypes of gradient steepness, location, intensity, and variability. The data manipulation protocol proposed can be achieved in a simple and efficient manner using the statistical programming language R, opening paths to perform high-throughput spatiotemporal phenotyping of intracellular gradients in apically growing cells.

Towards perceptual image fusion: A novel two-layer framework

Recent studies in neuroscience indicate that the human visual system perceives regular and irregular contents separately: the former mainly illustrate primary visual information such as image structures; while the latter are generally messy and independent, and seem to be less important in perception. However, without any reference to such perceptual theory, the existing image fusion algorithms treat these two types of contents equally, and may not preserve the most perceptually important information in source images. In this work, we propose a new two-layer image fusion framework towards consistency with human perception. The main contributions are as follows: (1) We firstly explore the recently revealed perceptual theory and characterize perceptual significance from regular and irregular image contents in image fusion. This creates the possibilities to develop fusion algorithms towards consistency with human perception and preserve more desirable information in the fused result. (2) Following the concept of active inference mechanisms in perception, we present a perceptual image decomposition model with a local regression method to separate images into regular and irregular layers. In this way, we could treat these two kinds of image contents discriminatively, with elaborately selected fusion strategies based on sparse representation and local energy. We conduct extensive experiments including subjective evaluation, objective evaluation and perceptual assessment, and the experimental results demonstrate the superiority of the proposed model.

LDV measurements of particle fluctuation velocities in dilute gravity-driven gas-particle flows

Particle instantaneous velocities were measured by a laser-Doppler velocimeter (LDV) in a vertical channel, where particles were supplied from the top of the channel and gas superficial velocity was close to zero. According to kinetic theory of granular flows, particle instantaneous velocities are decomposed into two components, a local Eulerian velocity that represents bulk motion of neighboring particles and a residual component, by three different local averaged methods. Results indicate that the local averaged method based on inter-particle collision time fails to accurately estimate the rapid variation of local Eulerian velocity. In contrast, the local Eulerian velocity estimated by both the simple moving averaged method and the robust local regression method could well represent the collective motions of neighboring particles. Moreover, the magnitude of the residual component (i.e., granular temperature) estimated by the robust local regression method was less sensitive to the number of neighboring data points used in the decomposition process. Overall trends of granular temperature and particle-phase turbulent kinetic energy in the wall normal direction are different from each other, indicating that the granular temperature and particle-phase turbulent kinetic energy must be treated separately to accurately predict energy balance. The probability density function of the residual component is compared with a Gaussian distribution. At the center of the channel, it follows a Gaussian distribution, while in the vicinity of the wall it retains a long tail.

Local regressions for decomposing CO2 and CH4 time-series in a semi-arid ecosystem

Local regressions have been widely employed for decomposing atmospheric data series. However, the use of local quadratic regressions is less extended. The current paper is grounded on the hypothesis that local linear regressions are able to capture CO2 and CH4 temporal evolution equally as well as quadratic linear regressions. Thus, the current paper pursues the following goals: (1) to quantify the temporal variability of both gases by using the local linear and local quadratic regression method; (2) to analyse the main statistics of the detrended series over time; (3) to compare results between the local linear and local quadratic regression method. Minimum mixing ratios in late summer and maximum in winter were detected for both gases. Atmospheric increases of an average of 1.98 ppm year−1 for CO2 and 11 ppb year−1 for CH4 were found when applying the local linear regression method. Alternatively, an increase of 2.24 ppm year−1 for CO2 and around 10.34 ppb year−1 for CH4 was obtained when the local quadratic regression method was applied. The Pearson correlation coefficients (0.21–0.40) showed acceptable values due to the large amount of available data. Statistically significant differences for the initial and the smoothed trend, as well as statistically significant differences for the seasonal component, were reported when comparing the local linear with the local quadratic method. Overall, both methods proved easy to apply and both provided acceptable data accuracy.

Research on the ecological compensation standard of the basin pollution control project based on evolutionary game theory and by taking Xiangjiang River as a case

Ecological compensation is a new resource and environment management model. As one of the main areas for implementing ecological compensation policies, basin ecological compensation has become an important measure for encouraging basin pollution control projects and improving the quality of regional economic development. By applying the basic game analysis of evolutionary game theory and building an evolutionary game model with a “reward—punishment” mechanism, this paper compares the interest-related decision-making behaviors of the upstream and downstream stakeholders of basin ecological compensation. By using data on the water quality of Xiangjiang River Basin, this paper calculates the rewards and penalties in different intervals by building a parametric regression mathematical model and employing the local linear regression method. Results show that a decline in water quality should be fined RMB 925500 yuan, an improvement in water quality should be awarded RMB 1227800 yuan, and a deteriorating water quality should be severely fined RMB 5087600 yuan.

A Photometric Redshift Catalog Based on SCUSS, SDSS, and WISE Surveys

Abstract We publish a photometric redshift catalog based on imaging data of the South Galactic Cap u-band Sky Survey, Sloan Digital Sky Survey (SDSS), and Wide-field Infrared Survey Explorer. A total of seven photometric bands are used, ranging from near-ultraviolet to near-infrared. A local linear regression method is adopted to estimate the photometric redshift with a dedicated spectroscopic training set. The photometric redshift catalog contains about 23.1 million galaxies classified by SDSS. Using the training set with redshift up to 0.8 and r-band magnitude down to 22 mag, we achieve an average bias of , a standard deviation of σ(Δz norm) = 0.019, and a 3σ outlier rate of about 4.2%. The bias is less than 0.01 at z &lt; 0.6 and goes up to about 0.05 at z ∼ 0.8. Compared with SDSS photometric redshifts, our redshift estimations are more accurate and have less bias.

Examining empirical evidence of the effect of superfluidity on the fusion barrier

Background: Several Time-Dependent Hartree-Fock-Bogoliubov (TDHFB) calculations predict that the super- fluidity enhances the fluctuations of the fusion barrier. This effect is not fully understood and not yet revealed experimentally. Purpose: The goal of this study is to investigate empirically the effect of the superfluidity on the fusion barrier width. Method: First, the local regression method is introduce and used to determine the barrier distribution more precisely. A second method that requires only the calculation of an integral of the cross section is developed to determine accurately the fluctuations of the barrier. A benchmark is done between this two methods and with the fitting method usually used. This integral method showing a better agreement in a test case, it is applied systematically in a selection of 115 fusion reactions. Results: The fluctuations of the barrier for superfluid systems are on average larger than for magic or semi-magic nuclei. This is due to the deformation effects and the effect of the superfluidity. To disentangle those two effects, we compare the experimental width to the width estimated from a model that takes into account the tunneling, the deformation and the vibration effect. The deviation of the experimental width from this theory for reaction between superfluid nuclei shows that the superfluidity enhance the fusion barrier width. Conclusions: This analysis shows that the predicted effect of the superfluidity on the width of the barrier is real and is of the order of 1 MeV.

Local linear estimation for spatial random processes with stochastic trend and stationary noise.

We consider the problem of estimating the trend for a spatial random process model expressed as Z(x) = μ(x) + ε(x) + δ(x), where the trend μ is a smooth random function, ε(x) is a mean zero, stationary random process, and {δ(x)} are assumed to be i.i.d. noise with zero mean. We propose a new model for stochastic trend in by generalizing the notion of a structural model for trend in time series. We estimate the stochastic trend nonparametrically using a local linear regression method and derive the asymptotic mean squared error of the trend estimate under the proposed model for trend. Our results show that the asymptotic mean squared error for the stochastic trend is of the same order of magnitude as that of a deterministic trend of comparable complexity. This result suggests from the point of view of estimation under stationary noise, it is immaterial whether the trend is treated as deterministic or stochastic. Moreover, we show that the rate of convergence of the estimator is determined by the degree of decay of the correlation function of the stationary process ε(x) and this rate can be different from the usual rate of convergence found in the literature on nonparametric function estimation. We also propose a data dependent selection procedure for the bandwidth parameter which is based on a generalization of Mallow's C p criterion. We illustrate the methodology by simulation studies and by analyzing a data on surface temperature anomalies.

Color Channel Reconstruction for Multi-Color Multi-View Images Using Disparity and Color Similarity-based Local Linear Regression

Most digital cameras today employ Bayer Color Filter Arrays in front of the camera sensor. In order to create a true-color image, a demosaicing step is required introducing image blur and artifacts. Special sensors like the Foveon X3 circumvent the demosaicing challenge by using pixels lying on top of each other. However, they are not commonly used due to high production cost and low flexibility. In this work, a multi-color multi-view approach is presented in order to create true-color images. Therefore, the red-filtered left view and the blue-filtered right view are registered and projected onto the green-filtered center view. Due to the camera offset and slightly different viewing angles of the scene, object occlusions might occur for the side channels, hence requiring the reconstruction of missing information. For that, a novel local linear regression method is proposed, based on disparity and color similarity. Simulation results show that the proposed method outperforms existing reconstruction techniques by on average 5 dB.

On Local Linear Regression Estimation of Finite Population Totals in Model Based Surveys

In this paper, nonparametric regression is employed which provides an estimation of unknown finite population totals. A robust estimator of finite population totals in model based inference is constructed using the procedure of local linear regression. In particular, robustness properties of the proposed estimator are derived and a brief comparison between the performances of the derived estimator and some existing estimators is made in terms of bias, MSE and relative efficiency. Results indicate that the local linear regression estimator is more efficient and performing better than the Horvitz-Thompson and Dorfman estimators, regardless of whether the model is specified or mispecified. The local linear regression estimator also outperforms the linear regression estimator in all the populations except when the population is linear. The confidence intervals generated by the model based local linear regression method are much tighter than those generated by the design based Horvitz-Thompson method. Generally the model based approach outperforms the design based approach regardless of whether the underlying model is correctly specified or not but that effect decreases as the model variance increases.

A Local Linear Regression Method Using a Discrete Kernel Function with Applications to Bond Curve Construction

Local linear regression is generally used to estimate the dependency of a random variable Y on another random variable X from a finite sample of data points. Except for the chosen kernel function, no model assumptions on the relationship between X and Y are taken into account in the estimation. In this paper, we show that discretizing the kernel function is consistent with a piecewise linear model assumption on the relationship between X and Y . Under certain boundary conditions, this relationship can be shown to be continuous. We apply the local linear regression with discretized kernel (LLRDK) to a problem from finance: the construction of a bond spread curve from a discrete set of observed bond market prices. The method we present comprises a valuation approach that combines a reference interest rate curve, survival probabilities from CDS markets and the statistical estimation of the bond CDS basis spreads using the LLRDK. This allows for the consistent combination of local linear regression as a statistical estimation method with model assumptions on the bond CDS spread such as a piecewise constant or piecewise linear dependency on time.

A sludge volume index (SVI) model based on the multivariate local quadratic polynomial regression method

Abstract In this study, a multivariate local quadratic polynomial regression (MLQPR) method is proposed to design a model for the sludge volume index (SVI). In MLQPR, a quadratic polynomial regression function is established to describe the relationship between SVI and the relative variables, and the important terms of the quadratic polynomial regression function are determined by the significant test of the corresponding coefficients. Moreover, a local estimation method is introduced to adjust the weights of the quadratic polynomial regression function to improve the model accuracy. Finally, the proposed method is applied to predict the SVI values in a real wastewater treatment process (WWTP). The experimental results demonstrate that the proposed MLQPR method has faster testing speed and more accurate results than some existing methods.

KNN-based local linear regression for the analysis and simulation of low flow extremes under climatic influence

Climate change frequently causes highly nonlinear and irregular behaviors in hydroclimatic systems. The stochastic simulation of hydroclimatic variables reproduces such irregular behaviors and is beneficial for assessing their impact on other regimes. The objective of the current study is to propose a novel method, a k-nearest neighbor (KNN) based on the local linear regression method (KLR), to reproduce nonlinear and heteroscedastic relations in hydroclimatic variables. The proposed model was validated with a nonlinear, heteroscedastic, lag-1 time dependent test function. The validation results of the test function show that the key statistics, nonlinear dependence, and heteroscedascity of the test data are reproduced well by the KLR model. In contrast, a traditional resampling technique, KNN resampling (KNNR), shows some biases with respect to key statistics, such as the variance and lag-1 correlation. Furthermore, the proposed KLR model was used to simulate the annual minimum of the consecutive 7-day average daily mean flow (Min7D) of the Romaine River, Quebec. The observed and extended North Atlantic Oscillation (NAO) index is incorporated into the model. The case study results of the observed period illustrate that the KLR model sufficiently reproduced key statistics and the nonlinear heteroscedasticity relation. For the future period, a lower mean is observed, which indicates that drier conditions other than normal might be expected in the next decade in the Romaine River. Overall, it is concluded that the KLR model can be a good alternative for simulating irregular and nonlinear behaviors in hydroclimatic variables.

Exploring Spatial and Temporal Distribution of Cutaneous Leishmaniasis in the Americas, 2001-2011.

Leishmaniasis is an important health problem in several countries in the Americas and cases notification is limited and underreported. In 2008, the Pan American Health Organization (PAHO/WHO) met with endemic countries to discuss the status and need of improvement of systems region-wide. The objective is to describe the temporal and spatial distribution of cutaneous leishmaniasis (CL) cases reported to PAHO/WHO by the endemic countries between 2001 and 2011 in the Americas.MethodsCases reported in the period of 2001–2011 from 14/18 CL endemic countries were included in this study by using two spreadsheet to collect the data. Two indicators were analyzed: CL cases and incidence rate. The local regression method was used to analyze case trends and incidence rates for all the studied period, and for 2011 the spatial distribution of each indicator was analyzed by quartile and stratified into four groups.ResultsFrom 2001–2011, 636,683 CL cases were reported by 14 countries and with an increase of 30% of the reported cases. The average incidence rate in the Americas was 15.89/100,000 inhabitants. In 2011, 15 countries reported cases in 180 from a total of 292 units of first subnational level. The global incidence rate for all countries was 17.42 cases per 100,000 inhabitants; while in 180 administrative units at the first subnational level, the average incidence rate was 57.52/100,000 inhabitants. Nicaragua and Panama had the highest incidence but more cases occurred in Brazil and Colombia. Spatial distribution was heterogeneous for each indicator, and when analyzed in different administrative level. The results showed different distribution patterns, illustrating the limitation of the use of individual indicators and the need to classify higher-risk areas in order to prioritize the actions. This study shows the epidemiological patterns using secondary data and the importance of using multiple indicators to define and characterize smaller territorial units for surveillance and control of leishmaniasis.

Prediction of soil salinity in the Yellow River Delta using geographically weighted regression

ABSTRACTIt is essential to determine the content and spatial distribution of soil salinity in a timely manner because soil salinization can cause land degradation on a regional scale. Geographically weighted regression (GWR) is a local regression method that can achieve the spatial extension of dependent variables based on the relationships between the dependent variables and environment variables and the spatial distances between the sample points and predicted locations. This study aimed to explore the feasibility of GWR in predicting soil salinity because the existing interpolation methods for soil salinity in the Yellow River Delta are still of low precision. Additionally, multiple linear regressions, cokriging and regression kriging were added to compare the accuracy of GWRs. The results showed that GWR predicted soil salinity with high accuracy. Furthermore, the accuracy was improved when compared to other methods. The root mean square error, correlation coefficient, regression coefficient and adjustment coefficients between the observed values and predicted values of the validation points were 0.31, 0.65, 0.57 and 0.42, respectively, which were better than that of other methods, indicating that GWR is an optimal method.