Mixture of Partial Least Squares Experts and Application in Prediction Settings with Multiple Operating Modes
نویسندگان
چکیده
This paper addresses the problem of online quality prediction in processes with multiple operating modes. The paper proposes a new method called mixture of partial least squares regression (Mix-PLS), where the solution of the mixture of experts regression is performed using the partial least squares (PLS) algorithm. The PLS is used to tune the model experts and the gate parameters. The solution of Mix-PLS is achieved using the expectationmaximization (EM) algorithm, and at each iteration of EM algorithm the number of latent variables of the PLS for the gate and experts are determined using the Bayesian information criterion. The proposed method, shows to be less prone to overfitting with respect to the number of mixture models, when compared to the standard R2Q1mixture of linear regression experts (MLRE). The Mix-PLS was successfully applied on three real prediction problems. The results were compared with five other regression algorithms. In all the experiments, the proposed method always exhibits the best prediction performance.
منابع مشابه
Removal of Brilliant Green and Crystal violet from Mono- and Bi-component Aqueous Solutions Using NaOH-modified Walnut Shell
In the present work, the simultaneous determination of Brilliant green (BG) and Crystal violet (CV) dyes with overlapped absorption spectra in binary mixture solution, was carreid out using the partial least squares (PLS) and direct ortogonal signal correction-partial least squares (DOSC-PLS) methods. The results obtained indicate that by applying DOSC on the calibration and prediction data for...
متن کاملLeast-squares support vector machine and its application in the simultaneous quantitative spectrophotometric determination of pharmaceutical ternary mixture
This paper proposes the least-squares support vector machine (LS-SVM) as an intelligent method applied on absorption spectra for the simultaneous determination of paracetamol (PCT), caffeine (CAF) and ibuprofen (IB) in Novafen. The signal to noise ratio (S/N) increased. Also, In the LS - SVM model, Kernel parameter (σ2) and capacity factor (C) were optimized. Excellent prediction was shown usin...
متن کاملRapid and Simultaneous Determination of Montelukast, Fexofenadine and Cetirizine Using Partial Least Squares and Artificial Neural Networks Modeling
Simultaneous determination of pharmaceutical compounds and accurate quantitative prediction of them are of great interest in the clinical and laboratory-based investigations.This work has focused on a comprehensive comparison of Partial Least-Squares (PLS-1) and Artificial Neural Networks (ANN) as two powerful types of chemometric methods. For this purpose, montelukast (MONT), fexofenadine ...
متن کاملSimultaneous spectrophotometric determination of lead, copper and nickel using xylenol orange by partial least squares
A partial least squares (PLS) calibration model was developed for the simultaneous spectrophotometricdetermination of Pb (ΙΙ), Cu (ΙΙ) and Ni (ΙΙ) using xylenol orange as a chromogenic reagent. The parameterscontrolling behavior of the system were investigated and optimum conditions were selected. The calibrationgraphs were linear in the ranges of 0.0–9.091, 0.0–2.719 and 0.0–2.381 ppm for lead...
متن کاملSimplex design method in simultaneous spectrophotometric determination of silicate and phosphate in boiler water of power plant and sewage sample by partial least squares
Partial least squares modeling as a powerful multivariate statistical tool was applied tothe simultaneous spectrophotometric determination of silicate and phosphate in aqueoussolutions. The concentration range for silicate and phosphate were 0.02-0.6 and 0.4-3 μg ml-1,respectively. The experimental calibration set was composed with 30 sample solutions using amixture design for two component mix...
متن کامل