Multivariate Adaptative Regression Splines (MARS), una alternativa para el análisis de series de tiempo

López, Jairo Vanegas; Vásquez, Fabián

doi:10.1016/j.gaceta.2016.10.003

Cited by 13 publications

(15 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One of the most widely used algorithms for solving adaptive computing problems is MARS [ 56 ]. This method consists of approximating an unknown function by the linear combination of a set of basic functions (products of the model variables) [ 57 ]. Among the key points of the algorithm, it stands out that it autonomously selects the relevant variables and interactions between them for each subregion.…”

Section: Methodsmentioning

confidence: 99%

Corrosion Prediction of Weathered Galvanised Structures Using Machine Learning Techniques

et al. 2021

View full text Add to dashboard Cite

Galvanised steel atmospheric corrosion is a complex multifactorial phenomenon that globally affects many structures, equipment, and sectors. Moreover, the International Organization of Standardization (ISO) standards require specific pollutant depositions values for any atmosphere classification or corrosion loss prediction result. The aim of this research is to develop predictive models to estimate corrosion loss based on easily worldwide available parameters. Experimental data from internationally validated studies were used for the data mining process, basing their characterisation on seven globally accessible qualitative and quantitative variables. Self-Organising Maps including both supervised and unsupervised layers were used to predict first-year corrosion loss, its corrosivity categories, and an uncertainty range. Additionally, a formula optimised with Newton’s method has been proposed for extrapolating these results to long-term results. The predictions obtained were compared with real values using Euclidean distances to know its similarity degree, offering high prediction performance. Specifically, evaluation results showed an average saving of up to 16% in coatings using these predictions. Therefore, using the proposed models reduces the uncertainty of the final structures state by predicting their material loss, avoiding initial over-dimensioning of structures, and meeting the principles of efficiency and sustainability, thus reducing costs.

show abstract

Section: Methodsmentioning

confidence: 99%

Corrosion Prediction of Weathered Galvanised Structures Using Machine Learning Techniques

et al. 2021

View full text Add to dashboard Cite

show abstract

“…First, the recursive partitioning algorithm and second, the MARS algorithm. The MARS algorithm includes two stepwise procedures to be performed in order to estimate and predict the model equation components, MARS‐forward and MARS‐backward stepwise 39 . Figure 5A,B shows the schematic view of the MARS model building procedure for piecewise linear basis functions with constant function B m ( x ) = 1.…”

Section: Methodology and Experimentationmentioning

confidence: 99%

“…Although, MARS method has been successfully applied 24‐26,32‐39 as one of the most prominent method for grading and regression in highly nonlinear systems, limited work has been reported in the field of battery SoC prediction due to problem associated with pruned data. The pruned data are generally not included in model development 40 .…”

Section: Introductionmentioning

confidence: 99%

State‐of‐charge prediction of lithium ion battery through multivariate adaptive recursive spline and principal component analysis

et al. 2020

View full text Add to dashboard Cite

The main aim of this research work is to provide a comprehensible state of art for the intensification of the utmost decisive task performed by a modern BMS system to monitor and estimate battery states through a well‐entrenched statistical analysis method. In the present work, “multivariate adaptive regression splines” (MARS) method along with principal component analysis (PCA) has been used to develop a predictive model‐based state of charge (SoC) estimator for an NCR 18650PF Lithium ion battery at constant charging c‐rate of 0.3 C and 0.3 C and 0.5 C constant discharge profiles. Time‐weighing factors, that is, voltage‐current and temperature are employed as training datasets, to provide greater impact for developing a SoC MARS model of with high coefficient of correlation R2 (0.9984). The SoC MARS model adequacy is then validated for voltage prediction of the same battery for two different profiles of discharging using NIPALS algorithm for principal component analysis (PCA) with SS2 of 93.69% and 94.23% for profile A and profile B, respectively.

show abstract

“…In each region in which the space is divided, a base linear function of one variable is adjusted. The final model is constituted from a combination of the generated base functions [53].…”

Section: Multivariate Adaptative Regression Splines (Mars)mentioning

confidence: 99%

Multivariate Analysis to Relate CTOD Values with Material Properties in Steel Welded Joints for the Offshore Wind Power Industry

et al. 2019

View full text Add to dashboard Cite

The increasingly mechanical requirements of offshore structures have established the relevance of fracture mechanics-based quality control in welded joints. For this purpose, crack tip opening displacement (CTOD) at a given distance from the crack tip has been considered one of the most suited parameters for modeling and control of crack growth, and it is broadly used at the industrial level. We have modeled, through multivariate analysis techniques, the relationships among CTOD values and other material properties (such as hardness, chemical composition, toughness, and microstructural morphology) in high-thickness offshore steel welded joints. In order to create this model, hundreds of tests were done on 72 real samples, which were welded with a wide range of real industrial parameters. The obtained results were processed and evaluated with different multivariate techniques, and we established the significance of all the chosen explanatory variables and the good predictive capability of the CTOD tests within the limits of the experimental variation. By establishing the use of this model, significant savings can be achieved in the manufacturing of wind generators, as CTOD tests are more expensive and complex than the proposed alternatives. Additionally, this model allows for some technical conclusions.

show abstract

Multivariate Adaptative Regression Splines (MARS), una alternativa para el análisis de series de tiempo

Cited by 13 publications

References 4 publications

Corrosion Prediction of Weathered Galvanised Structures Using Machine Learning Techniques

Corrosion Prediction of Weathered Galvanised Structures Using Machine Learning Techniques

State‐of‐charge prediction of lithium ion battery through multivariate adaptive recursive spline and principal component analysis

Multivariate Analysis to Relate CTOD Values with Material Properties in Steel Welded Joints for the Offshore Wind Power Industry

Contact Info

Product

Resources

About