Multivariate Adaptive Regression Splines Approach Research Articles

A New Predictive Model for the State-of-Charge of a High-Power Lithium-Ion Cell Based on a PSO-Optimized Multivariate Adaptive Regression Spline Approach

Batteries play a key role in achieving the target of universal access to reliable affordable energy. Despite their relevant importance, many challenges remain unsolved with regard to the characterization and management of batteries. One of the major issues in any battery application is the estimation of the state-of-charge (SoC). SoC, which is expressed as a percentage, indicates the amount of energy available in a battery. An accurate SoC estimation under realistic conditions improves battery performance, reliability, and lifetime. This paper proposes an SoC estimation method based on a new hybrid model that combines multivariate adaptive regression splines (MARS) and particle swarm optimization (PSO). The proposed hybrid PSO–MARS-based model uses data obtained from a high-power load profile (dynamic stress test) specified by the United States Advanced Battery Consortium (USABC). The results provide comparable accuracy to other more sophisticated techniques but at a lower computational cost.

A New Predictive Model Based on the ABC Optimized Multivariate Adaptive Regression Splines Approach for Predicting the Remaining Useful Life in Aircraft Engines

Remaining useful life (RUL) estimation is considered as one of the most central points in the prognostics and health management (PHM). The present paper describes a nonlinear hybrid ABC–MARS-based model for the prediction of the remaining useful life of aircraft engines. Indeed, it is well-known that an accurate RUL estimation allows failure prevention in a more controllable way so that the effective maintenance can be carried out in appropriate time to correct impending faults. The proposed hybrid model combines multivariate adaptive regression splines (MARS), which have been successfully adopted for regression problems, with the artificial bee colony (ABC) technique. This optimization technique involves parameter setting in the MARS training procedure, which significantly influences the regression accuracy. However, its use in reliability applications has not yet been widely explored. Bearing this in mind, remaining useful life values have been predicted here by using the hybrid ABC–MARS-based model from the remaining measured parameters (input variables) for aircraft engines with success. A correlation coefficient equal to 0.92 was obtained when this hybrid ABC–MARS-based model was applied to experimental data. The agreement of this model with experimental data confirmed its good performance. The main advantage of this predictive model is that it does not require information about the previous operation states of the aircraft engine.

A New Predictive Model of Centerline Segregation in Continuous Cast Steel Slabs by Using Multivariate Adaptive Regression Splines Approach

The aim of this study was to obtain a predictive model able to perform an early detection of central segregation severity in continuous cast steel slabs. Segregation in steel cast products is an internal defect that can be very harmful when slabs are rolled in heavy plate mills. In this research work, the central segregation was studied with success using the data mining methodology based on multivariate adaptive regression splines (MARS) technique. For this purpose, the most important physical-chemical parameters are considered. The results of the present study are two-fold. In the first place, the significance of each physical-chemical variable on the segregation is presented through the model. Second, a model for forecasting segregation is obtained. Regression with optimal hyperparameters was performed and coefficients of determination equal to 0.93 for continuity factor estimation and 0.95 for average width were obtained when the MARS technique was applied to the experimental dataset, respectively. The agreement between experimental data and the model confirmed the good performance of the latter.

Estimation of scour depth below free overfall spillways using multivariate adaptive regression splines and artificial neural networks

Erosion and scouring caused by the outflow jets of hydraulic structures is one of the most important topics in hydraulic engineering. In free overfall spillways, a water jet impacts the erodible downstream bed almost vertically and creates a scour hole. The scour hole can affect the safety and stability of the dam. In the present paper, the multivariate adaptive regression splines (MARS) approach has been adopted as a new soft computing tool for estimating the equilibrium scour depth below free overfall spillways. Using experimental data and dimensionless parameters, the MARS model has been developed to predict scour depth. Results obtained from the MARS approach were compared with those from the artificial neural network (ANN) and decision-tree algorithm. Statistical indicators demonstrated that the MARS approach had a good performance and presented competitive results, slightly better than ANN, for the prediction of this phenomenon.

MARSpline model for lead seven-day maximum and minimum air temperature prediction in Chennai, India

In this study, a Multivariate Adaptive Regression Spline (MARS) based lead seven days minimum and maximum surface air temperature prediction system is modelled for station Chennai, India. To emphasize the effectiveness of the proposed system, comparison is made with the models created using statistical learning technique Support Vector Machine Regression (SVMr). The analysis highlights that prediction accuracy of MARS models for minimum temperature forecast are promising for short term forecast (lead days 1 to 3) with mean absolute error (MAE) less than 1 °C and the prediction efficiency and skill degrades in medium term forecast (lead days 4 to 7) with slightly above 1 °C. The MAE of maximum temperature is little higher than minimum temperature forecast varying from 0.87 °C for day-one to 1.27 °C for lag day-seven with MARS approach. The statistical error analysis emphasizes that MARS models perform well with an average 0.2 °C of reduction in MAE over SVMr models for all ahead seven days and provide significant guidance for the prediction of temperature event. The study also suggests that the correlation between the atmospheric parameters used as predictors and the temperature event decreases as the lag increases with both approaches.

Statistical models for shear strength of RC beam‐column joints using machine‐learning techniques

SUMMARYThis paper proposes a new set of probabilistic joint shear strength models using the conventional multiple linear regression method, and advanced machine‐learning methods of multivariate adaptive regression splines (MARS) and symbolic regression (SR). In order to achieve high‐fidelity regression models with reduced model errors and bias, this study constructs extensive experimental databases for reinforced and unreinforced concrete joints by collecting existing beam‐column joint subassemblage tests from multiple sources. Various influential parameters that affect joint shear strength such as material properties, design parameters, and joint configuration are investigated through tests of statistical significance. After performing a set of regression analyses, the comparison of simulation results indicates that MARS approach is the best estimation method. Moreover, the accuracy of analytical predictions of the derived MARS model is compared with that of existing joint shear strength relationships. The comparison results show that the proposed model is more accurate compared to existing relationships. This joint shear strength prediction model can be readily implemented into joint response models for evaluation of earthquake performance and inelastic responses of building frames. Copyright © 2014 John Wiley & Sons, Ltd.

A Multivariate Adaptive Regression Spline Approach for Prediction of Maximum Shear Modulus and Minimum Damping Ratio

This study uses multivariate adaptive regression spline (MARS) for determination of maximum shear modulus (G max ) and minimum damping ratio (ξ min ) of synthetic reinforced soil. MARS employs confining pressure (σ, psi), rubber (r, %) and sand (s, %) as input variables. The outputs of the MARS are G max and ξ min . The developed MARS gives equations for determination of G max and ξ min . The results of MARS have been compared with the adaptive neuro-fuzzy inference system (ANFIS), multi-layer perception (MLP) and multiple regression analysis method (MRM). A sensitivity analysis has been also carried out to determine the effect of each input variable on G max and ξ min . This study shows that the developed MARS is a robust model for prediction of G max and ξ min .

Determination of ultimate capacity of driven piles in cohesionless soil: A Multivariate Adaptive Regression Spline approach

SUMMARYThe determination of ultimate capacity (Q) of driven piles in cohesionless soil is an important task in geotechnical engineering. This article adopts Multivariate Adaptive Regression Spline (MARS) for prediction Q of driven piles in cohesionless soil. MARS uses length (L), angle of shear resistance of the soil around the shaft (ϕshaft), angle of shear resistance of the soil at the tip of the pile (ϕtip), area (A), and effective vertical stress at the tip of the pile as input variables. Q is the output of MARS. The results of MARS are compared with that of the Generalized Regression Neural Network model. An equation has been also presented based on the developed MARS. The results show the strong potential of MARS to be applied to geotechnical engineering as a regression tool. Copyright © 2011 John Wiley & Sons, Ltd.

Mars approach for global sensitivity analysis of differential equation models with applications to dynamics of influenza infection.

Differential equation models are widely used for the study of natural phenomena in many fields. The study usually involves unknown factors such as initial conditions and/or parameters. It is important to investigate the impact of unknown factors (parameters and initial conditions) on model outputs in order to better understand the system the model represents. Apportioning the uncertainty (variation) of output variables of a model according to the input factors is referred to as sensitivity analysis. In this paper, we focus on the global sensitivity analysis of ordinary differential equation (ODE) models over a time period using the multivariate adaptive regression spline (MARS) as a meta model based on the concept of the variance of conditional expectation (VCE). We suggest to evaluate the VCE analytically using the MARS model structure of univariate tensor-product functions which is more computationally efficient. Our simulation studies show that the MARS model approach performs very well and helps to significantly reduce the computational cost. We present an application example of sensitivity analysis of ODE models for influenza infection to further illustrate the usefulness of the proposed method.

Forecasting Tourism Demand Using Time Series, Artificial Neural Networks and Multivariate Adaptive Regression Splines:Evidence from Taiwan

In the past few decades, international tourism has grown rapidly and has become a very interesting topic in tourism research. Taiwan, acting as a citizen in the global community, improved traveling facilities, and governments’ strong promotion has drawn more and more visitors to visit Taiwan. This study tries to build the forecasting model of visitors to Taiwan using three commonly adopted ARIMA, artificial neural networks (ANNs), and multivariate adaptive regression splines (MARS). In order to evaluate the appropriateness of the proposed modeling approaches, the dataset of monthly visitors to Taiwan was used as the illustrative example. Analytic results demonstrated that ARIMA outperformed ANNs and MARS approaches in terms of RMSE, MAD, and MAPE and provided effective alternatives for forecasting tourism demand.

Data mining in real estate appraisal: a model tree and multivariate adaptive regression spline approach

In this paper we adopt two exploratory modelling techniques: Model Trees and Multivariate Adaptive Regression Splines. The objective is the building of two sale price prediction models in order to highlight possible market segments not detectable a priori . We show how these novel procedures can help to understand complex patterns and interactions among predictors in real estate appraisal.

The Application of MARS on Crashworthiness Improvement

Surrogate model produced by using multivariate adaptive regression spline (MARS) approach takes the form of an expansion in a set of basis functions which are selected from data. This approach is especially useful in the case of having no advance understanding of the parametric model. As for MARS procedure, the adaptive adjustment is used frequently to best fit the data optimally. In this paper, the adaptive adjustment is realized by using Matlab programming language and provides support for MARS procedure. Taking crashworthiness improvement for example, the program is applied to produce surrogate model for peak acceleration, and then the optimization is carried out based on this model. The results indicate that the surrogate model constructed by MARS approach can predict peak acceleration precisely and therefore can provide instruct for crashworthiness improvement.

Vulnerability of carbon storage in North American boreal forests to wildfires during the 21st century

AbstractThe boreal forest contains large reserves of carbon. Across this region, wildfires influence the temporal and spatial dynamics of carbon storage. In this study, we estimate fire emissions and changes in carbon storage for boreal North America over the 21st century. We use a gridded data set developed with a multivariate adaptive regression spline approach to determine how area burned varies each year with changing climatic and fuel moisture conditions. We apply the process‐based Terrestrial Ecosystem Model to evaluate the role of future fire on the carbon dynamics of boreal North America in the context of changing atmospheric carbon dioxide (CO2) concentration and climate in the A2 and B2 emissions scenarios of the CGCM2 global climate model. Relative to the last decade of the 20th century, decadal total carbon emissions from fire increase by 2.5–4.4 times by 2091–2100, depending on the climate scenario and assumptions about CO2 fertilization. Larger fire emissions occur with warmer climates or if CO2 fertilization is assumed to occur. Despite the increases in fire emissions, our simulations indicate that boreal North America will be a carbon sink over the 21st century if CO2 fertilization is assumed to occur in the future. In contrast, simulations excluding CO2 fertilization over the same period indicate that the region will change to a carbon source to the atmosphere, with the source being 2.1 times greater under the warmer A2 scenario than the B2 scenario. To improve estimates of wildfire on terrestrial carbon dynamics in boreal North America, future studies should incorporate the role of dynamic vegetation to represent more accurately post‐fire successional processes, incorporate fire severity parameters that change in time and space, account for human influences through increased fire suppression, and integrate the role of other disturbances and their interactions with future fire regime.

Using multivariate adaptive regression splines (MARS) to identify relationships between soil and corn (Zea mays L.) production properties

Over-application of agricultural fertilizers can contribute to degradation of surface water quality. Factors governing crop establishment and yield must be identified in order to efficiently manage N application rates in corn (Zea mays L.) production systems. Spatial data sets of corn establishment and grain yields, and soil physical and chemical parameters were obtained for two corn production systems on a poorly drained clay loam soil in eastern Ontario, Canada, during low yielding conditions in 2000. The multivariate adaptive regression splines (MARS) automated regression data mining method was used to determine the dominant factors affecting both crop establishment and yield from these data sets. The analysis using MARS suggests that soil water content and cone penetration resistance are more important than elevation and spring mineral soil N (NH4+ + NO3−) in predicting crop establishment and grain yield. The MARS approach proved to be a useful method for identifying relationships between potential yield-governing variables. It also helped elucidate potential cause and effect processes, and in so doing, helped identify areas within the field where soil physical parameters may have been more important than nutrients in governing corn yield. Key words: Mulitvariate adaptive regression splines, corn yield, cone penetration resistance, soil water content, soil N, topography

A new technique for using multivariate adaptive regression splines (MARS) in pavement roughness prediction

The paper presents the application of a new statistical technique, multivariate adaptive regression splines (MARS), to a flexible pavement roughness prediction model. MARS is a non-parametric function estimation technique that shows great promise for fitting non-linear multivariate functions. The MARS approach was used to develop a roughness equation, based on available input, and was able to identify the threshold values of each input and the most important variables contributing to the roughness equation. The MARS technique allows easy interpretation of the relative importance of pavement condition variables, environmental factors and traffic for the overall fit.

A new technique for using multivariate adaptive regression splines (MARS) in pavement roughness prediction

The paper presents the application of a new statistical technique, multivariate adaptive regression splines (MARS), to a flexible pavement roughness prediction model. MARS is a non-parametric function estimation technique that shows great promise for fitting non-linear multivariate functions. The MARS approach was used to develop a roughness equation, based on available input, and was able to identify the threshold values of each input and the most important variables contributing to the roughness equation. The MARS technique allows easy interpretation of the relative importance of pavement condition variables, environmental factors and traffic for the overall fit.

Statistical techniques for the classification of chromites in diamond exploration samples

A database of 1267 quality-screened major- and trace-element analyses of chromites (s.l.) from kimberlites, lamproites, ultramafic lamprophyres (UMLs) and crustal sources (‘greenstones’, including ophiolites, gabbros, basalts and komatiites) has been subjected to statistical analysis, in order to derive discriminants for use in diamond exploration. The techniques used included nearest-neighbour analysis, CART (classification and regression trees) and MARS (multivariate adaptive regression splines). The results show that both CART and MARS approaches can correctly discriminate kimberlite/lamproite chromites from UML/‘greenstone’ chromites at levels near 90%. Discrimination into the four classes separately is achieved at levels of ca. 80% by MARS and > 70% by CART; much of the misclassification is between the kimberlite and lamproite classes. These results probably represent the maximum probable level of discrimination on chemical criteria, given that ascending magmas may sample both mantle and crustal rocks. The CART approach produces a classification tree that requires no further computation to classify a given grain; the MARS approach requires the use of a simple software package. Tests on known samples illustrate the high level of accuracy of the methods in an exploration context, as well as the useful petrogenetic conclusions that can be drawn from some ‘misclassifications’.