MVC1_GUI: A MATLAB graphical user interface for first-order multivariate calibration. An upgrade including artificial neural networks modelling

GOICOECHEA, HÉCTOR C.; OLIVIERI, ALEJANDRO C.; CHIAPPINI, FABRICIO A.

CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS

ELSEVIER SCIENCE BV

Año: 2020 vol. 206 p. 104162 - 104162

0169-7439

In the present report, an upgrade of a MATLAB graphical user interface (GUI) toolbox for implementing first-order multivariate calibration models is presented. The new freely available Multivariate Calibration 1 (MVC1_GUI) incorporates new models and features that make it a very versatile tool for data processing. In addition to the linear models, i.e., principal component regression (PCR) and partial least-squares 1 (PLS-1), included in the earlier software version, PLS-2 and maximum likelihood principal component regression (MLPCR) are now available, together with two non-linear models based on two different types of artificial neural networks (ANN): feed-forward multi-layer network with radial basis functions (RBF) and multi-layer back-propagation perceptron (MLP). The toolbox accepts different input data formats, and incorporates many advanced pre-processing algorithms to improve prediction capabilities. The development and validation of each model and its subsequent application to unknown samples is straightforward, since it generates many different plots regarding model performance, including outlier detection. Prediction results are produced along with analytical figures of merit and standard errors calculated by uncertainty propagation.