An application of subagging for the improvement of prediction accuracy of multivariate calibration models

Variables selection methods in near-infrared spectroscopy

“…It is capable of calibrating mixtures with almost identical spectra without loss of prediction capacity using the spectrophotometric method. Many studies have demonstrated the importance of the validation step in GA wavelength selection for multiple linear regression in order to avoid random correlation and the selection of irrelevant variables [43,46,77,[79][80][81].…”

Bagging for robust non-linear multivariate calibration of spectroscopy

“…To better evaluate the advantage of bagging/subagging, a relative RMSEP reduction was also calculated as [37] ‫ۃ‬RMSEP ୧୬ୢ୧୴୧ୢ୳ୟ୪ ‫ۄ‬ − ‫ۃ‬RMSEP (ୱ୳)ୠୟ ୧୬ ‫ۄ‬…”

“…Among these modified methods, subagging was reported to provide similar performance to bagging with less computation, since it uses a subset of the data for model development [41]. For the purpose of multivariate spectroscopic calibration, subagging has already been applied to linear methods as PLS and MLR with variable selection [37]; yet its role in improving non-linear calibration methods (such as ANN and GPR) has been under-explored. Previous studies demonstrated that non-linear regression for spectroscopic calibration can achieve accurate prediction of analyte properties [15][16][17][18].…”

“…Subsequently, attention has been focused on bagging neural networks [31][32][33][34][35][36]. Recently, bagging was developed for other regression methods, such as partial least squares (PLS), multiple linear regression (MLR) [37], and Gaussian process regression (GPR) [38]. Over the years, bagging has been modified in several ways such as "nice" bagging [39], iterated bagging [40], subagging (sub-sample aggregating, based on sub-sampling without replacement) [41], and trimmed bagging [42].…”

Entropy-Based Bagging for Fault Prediction of Transformers Using Oil-Dissolved Gas Data

“…Adaboost is adapted to combine weak classification or prediction models with strong independence, while Bagging is applicable to combine strong classification or prediction models. Some approaches, such as sub-Bagging [11] and weighted member models [12], were brought forward to improve ensemble learning with satisfactory results. This paper presents an improved Bagging algorithm for power transformer fault prediction based on oil-dissolved gas.…”