Szczegóły publikacji
Opis bibliograficzny
Ridge regression with self - Paced learning algorithm in interpretation of voltammetric signals / Łukasz GÓRSKI, Małgorzata JAKUBOWSKA // Chemometrics and Intelligent Laboratory Systems ; ISSN 0169-7439. — 2019 — vol. 191, s. 73–81. — Bibliogr. s. 81, Abstr. — Publikacja dostępna online od: 2019-06-25
Autorzy (2)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 122939 |
|---|---|
| Data dodania do BaDAP | 2019-07-22 |
| Tekst źródłowy | URL |
| DOI | 10.1016/j.chemolab.2019.06.008 |
| Rok publikacji | 2019 |
| Typ publikacji | artykuł w czasopiśmie |
| Otwarty dostęp |  |
| Czasopismo/seria | Chemometrics and Intelligent Laboratory Systems |
Abstract
Application of Ridge Regression (RR) with Self – Paced Learning (SPL) function in voltammetry was presented in this work. RR provided multivariate regression model with simple linear relation between dependent variables y (concentrations) and independent variables X (voltammograms). Additionally, SPL function introduced smart approach to model construction, by gradual samples incorporation to learning dataset, controlled by the parameter called model age k. Both, optimization of the model parameters and evaluation of its prediction abilities, was done by venetian blinds cross validation. The model with the smallest cross validation error (RMSECV) was considered as optimal. The results of RR-SPL were compared with PLS and PCR, using simulated data and experimental overlapping DPV signals of In(III), Tl(I) and Pb(II). Conjugation of RR with SPL provided calibration models with better prediction abilities (smaller cross validation errors), compared to the typical PLS and PCR approaches. © 2019 Elsevier B.V.
