Utilização de metodos quimioterapicos em dados de natureza multivariada

AUTOR(ES)
DATA DE PUBLICAÇÃO

2003

RESUMO

Chemometric methods of data analysis are applied in this work to experimental data provided by IAC (Instituto Agronômico de Campinas) and Henkel Chemical Industries. Firstly, a brief introduction of the quali and quantitatives multivariate methods, the concepts involved, their advantages and disadvantages, is presented with the goal to give a general idea on how the provided data will be treated. The first data set referees to the study of the artificial degreening process applied to Murcote tangerines. The fruits were treated with different gases, at different temperatures and chemical analyses were carried out during a time interval, with the goal to evaluate the efficiency of the treatments used. For this data set, the study was done using multivariate exploratory analysis (PCA and HCA), N-Way techniques (PARAFAC) and the analysis of variance (ANOVA). The methods were compared and their performance evaluated. The second data set consisted of absorbance spectra registered in the near infrared region from epoxide soybean oil samples. Industrially, it is necessary to quantify some analytes, present in the epoxide soybean oil, to control the final product quality. These analytes are the residual water percentage and the iodine and epoxide indices, these last two indicating the efficiency of the stabilizing properties of the product. In this way, these spectra were used in the construction of calibration models, saving time and money for chemical determination of the analytes. The multivariate regression method PLS was used, exploring some variables selection techniques as well.

ASSUNTO(S)

quimioterapia calibração multivariada

ACESSO AO ARTIGO

Documentos Relacionados