PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 11131897

Data evaluation for soft drink quality control using principal component analysis and back-propagation neural networks.

G González·E M Pena Méndez·M J Sánchez Sánchez·J Havel
View original source ↗Export metadata ↓

Abstract

This work describes an alternative for chemical data research, with the aim of evaluating finished product quality. Analytical data for additives in soft drinks are interpreted by the use of multivariate data analysis: principal component analysis (PCA), factor analysis, cluster analysis, and artificial neural networks. Taking into account various chemical components like sorbic, benzoic, and ascorbic acids; saccharose; caffeine; Na, K, Ca, Mg, Fe, Zn, Cu, P, and B, soft drinks were characterized and classified. The ratios of Na, K, Ca + Mg, P, and K/Na have been studied. The application of PCA, cluster analysis, and artificial neural networks showed that combination of these chemometric tools offers effective means for modeling and classifying soft drinks in accordance with their contents in additives and heavy metals.

Explore related subjects

BeveragesCluster AnalysisFactor Analysis, StatisticalFood AdditivesHydrogen-Ion ConcentrationModels, TheoreticalMultivariate AnalysisNeural Networks, ComputerQuality ControlJournal Article

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

G González, E M Pena Méndez, M J Sánchez Sánchez, J Havel. 2000. Data evaluation for soft drink quality control using principal component analysis and back-propagation neural networks.. https://doi.org/10.4315/0362-028x-63.12.1719

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Determination of isopropyl-9H-thioxanthen-9-one in packaged beverages by solid-phase extraction clean-up and liquid chromatography with tandem mass spectrometry detection.

A simple method was developed and validated for the trace determination of 2-isopropylthioxanthone (ITX) in packaged drinks. Samples were extracted from the food matrix using acetonitrile:water (60:40, v/v), and further subjected to clean-up and preconcentration using solid-phase extraction prior to analysis by liquid chromatography-tandem mass spectrometry using multiple reaction monitoring (MRM) mode. The use of 2-isopropyl-[(2)H7]thioxanthen-9-one was incorporated into the method as an internal standard. Excellent 3-day interday precision data (RSD 0.72%, n=10), and intraday precision data (RSD 0.52%, n=10) were obtained on a 0.10 microg/L standard solution. Spiked samples (n=8) were used to gauge the accuracy of the method at the concentration levels of 2.5, 100, and 500 microg/kg in food; recoveries ranged from 97.0 to 103.0%. These excellent validation data suggest the exciting possibility of using this method for the determination of low levels of ITX migrating from printed food packaging materials into beverages with a method quantitation limit of 0.50 microg/kg. For the first time, analysis on a range of milk, juice, tea and yoghurt drinks, as well as their respective food packaging materials were performed for comparative studies on their ITX content.

Beverages↗
PubMed2007

The effect of food simulants on porcelain-composite bonding.

PURPOSE: The purpose of this study was to evaluate the in vitro effects of food simulants on the bond strength of brackets to porcelain surfaces. MATERIALS AND METHODS: One hundred and eighty porcelain-fused to metal specimens were fabricated. Samples were divided into 3 groups including 60 specimens each. Different surface treatment methods were applied to each group and these were: sandblasting (SB), etching with orthophosphoric acid (OPA) and etching with hydrofluoric acid (HFA). After the surface treatments, the brackets were bonded to the center of each sample with a light-curing orthodontic composite adhesive. Then each group was divided into five test subgroups and one control subgroup each consisting of 10 specimens. The test groups were conditioned for 90 days at 37 degrees C as follows: water, 0.02N citric acid, heptane, 8% ethanol aqueous solution and 50% ethanol aqueous solution. The control specimens were stored at room temperature in air. Shear force was applied to the porcelain-bracket interface by using a universal testing machine with a crosshead speed of 1 mm/min until debonding occurred. The shear bond strength was calculated by dividing the maximum load by the cross-sectional area of the bracket to give the results in megapascals (MPa). The statistical evaluations were made by using analysis of variance. Whenever a significance was detected, Bonferroni tests were performed for post hoc analyses. RESULTS: Fifty percent ethanol conditioned specimens for all surface treatment groups showed lower shear bond strength values. HFA group showed significantly higher shear bond strength values when compared with other groups (P < 0.05). For all porcelain surface treatment groups, control groups showed statistically significant higher shear bond strength values (P < 0.05). CONCLUSION: The results of this in vitro study show that food simulants significantly decrease the bond strength between bracket and porcelain surface.

Beverages↗
PubMed2007

NADH screen-printed electrodes modified with zirconium phosphate, Meldola blue, and Reinecke salt. Application to the detection of glycerol by FIA.

Bulk screen-printed electrodes (bSPEs) modified with zirconium phosphate (ZrP) and Meldola blue (MB) and by electrochemical deposition of a Reineckate film (bMBZrPRs-SPEs) have been constructed and used as NADH sensors. Cyclic voltammetric investigation of these bulk electrochemically modified screen-printed electrodes revealed stable catalytic activity in oxidation of the reduced form of the coenzyme nicotinamide adenine dinucleotide (NADH). Flow-injection analysis (FIA) coupled with amperometric detection confirmed the improved stability of the bMBZrPRs-SPEs (10(-4) mol L(-1) NADH, %RSD = 4.2, n = 90, pH 7.0). Other conditions, for example applied working potential (+50 mV relative to Ag|AgCl), flow rate (0.30 mL min(-1)) and pH-dependence (range 4.0-10.0) were evaluated and optimized. A glycerol biosensor, prepared by immobilizing glycerol dehydrogenase (GDH) on the working electrode area of a bMBZrPRs-SPE, was also assembled. The biosensor was most stable at pH 8.5 (%RSD = 5.6, n = 70, 0.25 mmol L(-1) glycerol). The detection and quantification limits were 2.8 x 10(-6) and 9.4 x 10(-6) mol L(-1), respectively, and the linear working range was between 1.0 x 10(-5) and 1.0 x 10(-4) mol L(-1). To assess the effect of interferences, and recovery by the probe we analyzed samples taken during fermentation of chemically defined grape juice medium and compared the results with those obtained by HPLC.

Beverages↗