Ridge Regression Technique Research Articles

Inter-fuel substitution possibilities in South Africa: A translog production function approach

This study applies the translog production function to investigate technical change and energy substitution possibilities among petroleum, coal and electricity over the period 1980–2012. Ridge regression technique is introduced to correct for multicollinearity in the data. The study documents several findings: first, electricity and coal are found to be the major drivers of South African output and also have a faster technological progress over petroleum. Second, all energy inputs were found to be substitutes; therefore removing all price ceilings and subsidies on petroleum will decrease the demand for petroleum in effect protecting South African economy from external petroleum price shocks while reducing CO2 emissions. This will also increase the demand for electricity from renewable sources; however the success of this substitution will depend on policies geared towards large scale electricity production to meet demand. Third and finally, this study points to evidence that, even though coal dominates as the main energy source of South Africa, enhancement in research and development of renewable energy technologies could present opportunities for electricity as a potential replacer of coal; and as such, accelerating the CO2 mitigation effort of the South African government.

Machine learning approaches to environmental disturbance rejection in multi-axis optoelectronic force sensors

Light-intensity modulated (LIM) force sensors are seeing increasing interest in the field of surgical robotics and flexible systems in particular. However, such sensing modalities are notoriously susceptible to ambient effects such as temperature and environmental irradiance which can register as false force readings. We explore machine learning techniques to dynamically compensate for environmental biases that plague multi-axis optoelectronic force sensors. In this work, we fabricate a multisensor: three-axis LIM force sensor with integrated temperature and ambient irradiance sensing manufactured via a monolithic, origami-inspired fabrication process called printed-circuit MEMS. We explore machine learning regression techniques to compensate for temperature and ambient light sensitivity using on-board environmental sensor data. We compare batch-based ridge regression, kernelized regression and support vector techniques to baseline ordinary least-squares estimates to show that on-board environmental monitoring can substantially improve sensor force tracking performance and output stability under variable lighting and large (>100°C) thermal gradients. By augmenting the least-squares estimate with nonlinear functions describing both environmental disturbances and cross-axis coupling effects, we can reduce the error in Fx, Fy and Fz by 10%, 33%, and 73%, respectively. We assess viability of each algorithm tested in terms of both prediction accuracy and computational overhead, and analyze kernel-based regression for prediction in the context of online force feedback and haptics applications in surgical robotics. Finally, we suggest future work for fast approximation and prediction using stochastic, sparse kernel techniques.

Prediction of egg weight from egg quality characteristics via ridge regression and regression tree methods

This study was conducted on 2049 eggs, collected from commercial white layer hybrids, with the purpose of predicting egg weight (EW) from egg quality characteristics such as shell weight (SW), albumen weight (AW), and yolk weight (YW). In the prediction of EW, ridge regression (RR), multiple linear regression (MLR), and regression tree analysis (RTM) methods were used. Predictive performance of RR and MLR methods was evaluated using the determination coefficient (R2) and variance inflation factor (VIF). R2 (%) coefficients for RR and MLR methods were found as 93.15% and 93.4% without multicollinearity problems due to very low VIF values, varying from 1 to 2, respectively. Being a visual, non-parametric analysis technique, regression tree method (RTM) based on CHAID algorithm performed a very high predictive accuracy of 99.988% in the prediction of EW. The highest EW (71.963 g) was obtained from eggs with AW > 41 g and YW > 17 g. The usability of RTM due to a very great accuracy of 99.988 (%R2) in the prediction of EW could be advised in practice in comparison with the ridge regression and multiple linear regression analysis techniques, and might be a very valuable tool with respect to quality classification of eggs produced in the poultry science.

Reliability based design optimization of concrete mix proportions using generalized ridge regression model

This paper presents Reliability Based Design Optimization (RBDO) model to deal with uncertainties involved in concrete mix design process. The optimization problem is formulated in such a way that probabilistic concrete mix input parameters showing random characteristics are determined by minimizing the cost of concrete subjected to concrete compressive strength constraint for a given target reliability. Linear and quadratic models based on Ordinary Least Square Regression (OLSR), Traditional Ridge Regression (TRR) and Generalized Ridge Regression (GRR) techniques have been explored to select the best model to explicitly represent compressive strength of concrete. The RBDO model is solved by Sequential Optimization and Reliability Assessment (SORA) method using fully quadratic GRR model. Optimization results for a wide range of target compressive strength and reliability levels of 0.90, 0.95 and 0.99 have been reported. Also, safety factor based Deterministic Design Optimization (DDO) designs for each case are obtained. It has been observed that deterministic optimal designs are cost effective but proposed RBDO model gives improved design performance.

Kernel ridge regression for general noise model with its application

The classical ridge regression technique makes an assumption that the noise is Gaussian. However, it is reported that the noise models in some practical applications do not satisfy Gaussian distribution, such as wind speed prediction. In this case, the classical regression techniques are not optimal. So we derive an optimal loss function and construct a new framework of kernel ridge regression technique for general noise model (N-KRR). The Augmented Lagrangian Multiplier method is introduced to solve N-KRR. We test the proposed technique on artificial data and short-term wind speed prediction. Experimental results confirm the effectiveness of the proposed model.

Employing Ridge Regression Technique in Prediction

This paper is concerned with fitting some black box models. Some of them are, the outputs error model which contains the autoregressive and autoregressive moving average with additional inputs(ARX and ARMAX).The best model has been chosen which represents the data about the Temperature which is affected by some predictor variables which they were represented by (Brightness, unlike radiation, and evaporation).The parameters of the best model were estimated by the ridge regression method with and without the existence of prior information around the model parameters. The prediction errors of the model which has been estimated by least square and ridge regression when the prior information about the parameters is available were compared.

Employing Ridge Regression Technique in Prediction

Employing Ridge Regression Technique in Prediction

Linear Receiver for the Uplink in Distributed Antenna Systems

We consider the uplink of a distributed antenna system (DAS) in the presence of interference and unknown oscillator offsets, which lead to synchronization errors. For this scenario, we develop a linear receiver which maximizes the output signal-to-interference-plus-noise ratio (SINR). Specifically, we first propose a new structured generalized sidelobe canceller (SGSC) formulation for the commonly used minimum-variance receiver, and derive a general framework to improve the robustness of a linear receiver when oscillator offsets exist. Then a new linear receiver is proposed in closed-form based on the proposed robust SGSC and a ridge regression technique. It is shown by simulations that the proposed linear receiver can provide very robust SINR performance and excellent symbol error rate (SER) performance in the existence of interference and unknown oscillator offsets.

Feasibility of fuel cycle characterization using multiple nuclide signatures

The feasibility of identifying spent nuclear fuel arising from an unknown fuel cycle in terms of reactor type and burnup using a database of nuclide composition vectors generated for combinations of these two variables is examined. The database and test cases were generated using ORIGEN-ARP, and the concentrations of 200 nuclides were analyzed for each sample. Nearest neighbors and ridge regression techniques were used to make predictions of the reactor type and burnup of test cases. Various truncated nuclide lists were also tested. An initial examination of the techniques’ sensitivity to measurement error was made by perturbing the unknowns’ composition vector and examining the effect on each of the technique’s predictions. We demonstrate through the results of these experiments that investigation and development of multivariate data analysis methodologies for nuclear forensics applications is warranted.

Cascade correlation networks for electricity spot price forecasting in Brasil

The aim of this paper is to propose the use of regularized cascade correlation neural networks to forecast the monthly Brazilian electricity spot price. The cascade correlation models have been regularized with weight decay, weight elimination and ridge regression techniques, and several regularized models have been estimated. The results show that the regularized cascade correlation network represents the dynamic series better than other similar models such as the multilayer perceptron (MLP) and ARIMA. Then the regularized cascade correlation neural networks allow finding a suitable model to forecast the monthly Brazilian electricity spot price.

Partial Least Squares (PLS) Applied to Medical Bioinformatics

PLS initially creates uncorrelated latent variables which are linear combinations of the original input vectors Xi, where weights are used to determine linear combinations, which are proportional to the covariance. Secondly, a least squares regression is then performed on the subset of extracted latent variables that lead to a lower and biased variance on transformed data. This process, leads to a lower variance estimate of the regression coefficients when compared to the Ordinary Least Squares regression approach. Classical Principal Component Analysis (PCA), linear PLS and kernel ridge regression (KRR) techniques are well known shrinkage estimators designed to deal with multi- collinearity, which can be a serious problem. That is, multi-collinearity can dramatically influence the effectiveness of a regression model by changing the values and signs of estimated regression coefficients given different but similar data samples, thereby leading to a regression model which represents training data reasonably well, but generalizes poorly to validation and test data. We explain how to address these problems, which is followed by performing a PLS hypotheses driven preliminary research study and sensitivities analysis by not doing a combinatorial analysis as PLS will eliminate the unnecessary variables using a microarray colon cancer data set. Research studies as well as preliminary results are described in the results section.

Estimation of State Space Models by using Ridge Regression Technique with Application

In this paper many state space models are fitted. The statistical criteria used to choose the best estimated model whose parameters estimated by the Ridge regression method. This method can be used when the predicator variables are multicollinear.The data used in this paper represented by the temperature, which is the response variable and the predictor variables effulgence, wined, sunshine and evaporation.

Estimation of State Space Models by using Ridge Regression Technique with Application

Estimation of State Space Models by using Ridge Regression Technique with Application

Adaptive ridge regression system for software cost estimating on multi-collinear datasets

Cost estimation is one of the most critical activities in software life cycle. In past decades, a number of techniques have been proposed for cost estimation. Linear regression is yet the most frequently applied method in the literature. However, a number of studies point out that linear regression is prone to low prediction accuracy. The low prediction accuracy is due to a number of reasons such as non-linearity and non-normality. One less addressed reason is the multi-collinearities which may lead to unstable regression coefficients. On the other hand, it has been reported that multi-collinearity spreads widely across the software engineering datasets. To tackle this problem and improve regression's accuracy, we propose a holistic problem-solving approach (named adaptive ridge regression system) integrating data transformation, multi-collinearity diagnosis, ridge regression technique and multi-objective optimization. The proposed system is tested on two real world datasets with the comparisons with OLS regression, stepwise regression and other machine learning methods. The results indicate that adaptive ridge regression system can significantly improve the performance of regressions on multi-collinear datasets and produce more explainable results than machine learning methods.

Tensor-based projection using ridge regression and its application to action classification

In this study, the authors encode video sequence by a third-order tensor instead of matrices or vectors and present a novel framework of action classification via tensor-based projection using ridge regression (TPRR). The main contributions of this framework include: (i) Classification can be directly achieved by projecting the tensor samples into their binary class label through ridge regression technique. (ii) TPRR algorithm is further generalised from binary classification to solve multi-category classification problem by incorporating the error correcting concept. Experiments of action classification on a small or medium training set demonstrate that this approach outperforms previous tensor embedding methods and other state-of-the-art techniques, whereas on large training set, it is also able to achieve very satisfactory results. The proposed approach is further proven to be considerably robust to viewpoint, partial occlusion and irregularities in motion styles.

Asymptotic confidence intervals in ridge regression based on the Edgeworth expansion

Ridge Regression techniques have been found useful to reduce mean square errors of parameter estimates when multicollinearity is present. But the usefulness of the method rest not only upon its ability to produce good parameter estimates, with smaller mean squared error than Ordinary Least Squares, but also on having reasonable inferential procedures. The aim of this paper is to develop asymptotic confidence intervals for the model parameters based on Ridge Regression estimates and the Edgeworth expansion. Some simulation experiments are carried out to compare these confidence intervals with those obtained from the application of Ordinary Least Squares. Also, an example will be provided based on the well known data set of Hald.

Model-based monitoring and fault diagnosis of fossil power plant process units using Group Method of Data Handling

This paper presents an incipient fault diagnosis approach based on the Group Method of Data Handling (GMDH) technique. The GMDH algorithm provides a generic framework for characterizing the interrelationships among a set of process variables of fossil power plant sub-systems and is employed to generate estimates of important variables in a data-driven fashion. In this paper, ridge regression techniques are incorporated into the ordinary least squares (OLS) estimator to solve regression coefficients at each layer of the GMDH network. The fault diagnosis method is applied to feedwater heater leak detection with data from an operating coal-fired plant. The results demonstrate the proposed method is capable of providing an early warning to operators when a process fault or an equipment fault occurs in a fossil power plant.

Accommodating Linkage Disequilibrium in Genetic-Association Analyses via Ridge Regression

Large-scale genetic-association studies that take advantage of an extremely dense set of genetic markers have begun to produce very compelling statistical associations between multiple makers exhibiting strong linkage disequilibrium (LD) in a single genomic region and a phenotype of interest. However, the ultimate biological or "functional" significance of these multiple associations has been difficult to discern. In fact, the LD relationships between not only the markers found to be associated with the phenotype but also potential functionally or causally relevant genetic variations that reside near those markers have been exploited in such studies. Unfortunately, LD, especially strong LD, between variations at neighboring loci can make it difficult to distinguish the functionally relevant variations from nonfunctional variations. Although there are (rare) situations in which it is impossible to determine the independent phenotypic effects of variations in LD, there are strategies for accommodating LD between variations at different loci, and they can be used to tease out their independent effects on a phenotype. These strategies make it possible to differentiate potentially causative from noncausative variations. We describe one such approach involving ridge regression. We showcase the method by using both simulated and real data. Our results suggest that ridge regression and related techniques have the potential to distinguish causative from noncausative variations in association studies.

Ridge regression technique to determine the environmental influences on tef (Eragrostis tef) grain yield

Tef grain yield is dependent on a number of component characters such as plant height and panicle length. These characters and consequently yield are governed by a large number of factors including environmental factors. The objective of this paper was to determine the environmental influences on tef grain yield. The effects of eleven environmental variables on tef yield were studied using least squares and ridge regression analyses. The results revealed that the least squares estimates of regression coefficients for seven environmental variables did not give a correct indication of the influence of the variables on tef yield but that the estimates from ridge regression were stable. In addition, the ridge regression model has a lower mean square error. Tef grain yield is positively correlated to rainfall, average monthly minimum temperatures, soils dominated by silt with adequate quantities of available nitrogen while it is negatively influenced by average monthly maximum temperatures and clay soils. The results obtained from this study are in accordance with previous literature about the effects of environmental factors on yield of tef cultivars.

Automatic robust adaptive beamforming via ridge regression

In this paper we derive a class of new parameter free robust adaptive beamformers using the generalized sidelobe canceler reparameterization of the unit gain constrained minimum variance problem. In this parameterization the minimum variance beamformer is obtained as the solution of a linear least squares (LS) problem. In the case of an inaccurate steering vector and/or few data snapshots this marginally overdetermined system gives an ill fit causing signal cancellation in the standard minimum variance (LS) solution. By regularizing the LS problem using ridge regression techniques we get a whole class of robust adaptive beamformers, none of which requires the choice of a user parameter, as opposed to many existing methods. In this context we also propose a parameter free empirical Bayes-based ridge regression technique which, to the best of our knowledge, is novel. The performance of our approach is illustrated by numerical simulations and compared to other robust adaptive beamformers.