Premium
Application of partial least‐squares regression to the suitability of multicomponent polarographic determination of organochlorine pesticides in emulsified medium
Author(s) -
Reviejo A Julio,
Buyo Federico J.,
Pingarrón José M.,
Peral José L.
Publication year - 1993
Publication title -
electroanalysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.574
H-Index - 128
eISSN - 1521-4109
pISSN - 1040-0397
DOI - 10.1002/elan.1140050407
Subject(s) - calibration , partial least squares regression , chromatography , chemistry , chemometrics , analytical chemistry (journal) , heptachlor , matrix (chemical analysis) , statistics , mathematics , pesticide , dieldrin , agronomy , biology
Partial least‐squares regression (PLSR) has been used for treatment of the polarographic data obtained from mixtures of dieldrin, heptachlor, endosulfan, and endosulfan sulfate in an emulsified medium formed with ethyl acetate and a mixture of Triton X‐405 and Hyamine 2389 as emulsifying agent. Pesticide concentrations were within the range of 1.0 × 10 −6 to 9.0 × 10 −6 mol L −1 . The size of the calibration set was 35 samples by nine current measurements at different potentials. The overall „poorness” or „richness” of a sample can be predicted by plotting the samples in the space of the first two latent variables of the intensities matrix. The model was improved by constructing a new calibration set composed of samples containing only three and four components. In order to distinguish among the different types of samples (binary, ternary, etc.) a linear cutoff function was defined by dividing the sampling space into two calibration zones. Another assignment criterion for both calibration sets was the trunked Mahalanobis distances. External validation of the calibration sets was accomplished with seven new „unknown” samples. All of these samples were properly assigned using the cutoff linear function, and prediction of the pesticides concentration in each sample was possible with acceptable errors.