Goiano Federal Institute of Education

Brazil

Goiano Federal Institute of Education

Brazil
SEARCH FILTERS
Time filter
Source Type

De Souza Costa E.T.,Federal University of Lavras | Guilherme L.R.G.,Federal University of Lavras | De Melo E.E.C.,Federal University of Paraiba | Ribeiro B.T.,Federal University of Uberlandia | And 4 more authors.
Biological Trace Element Research | Year: 2012

This study evaluated Cd and Pb accumulation by castor bean (Ricinus communis cv. Guarany) plants grown in nutrient solution, aiming to assess the plant's ability and tolerance to grow in Cd- and Pb-contaminated solutions for phytoremediation purposes. The plants were grown in individual pots containing Hoagland and Arnon's nutrient solution with increasing concentrations of Cd and Pb. The production of root and shoot dry matter and their contents of Cd, Pb, Ca, Mg, Cu, Fe, Mn, and Zn were evaluated in order to calculate the translocation and bioaccumulation factors, as well as toxicity of Cd and Pb. Cadmium caused severe symptoms of phytotoxicity in the plant's root and shoot, but no adverse effect was observed for Pb. Castor bean is an appropriate plant to be used as indicator plant for Cd and tolerante for Pb in contaminated solution and it can be potentially used for phytoremediation of contaminated areas. © 2011 Springer Science+Business Media, LLC.


Ferreira J.C.,Goiano Federal Institute of Education | Vural E.,Middle East Technical University | Guillemot C.,French Institute for Research in Computer Science and Automation
IEEE Transactions on Image Processing | Year: 2016

Local learning of sparse image models has proved to be very effective to solve inverse problems in many computer vision applications. To learn such models, the data samples are often clustered using the K-means algorithm with the Euclidean distance as a dissimilarity metric. However, the Euclidean distance may not always be a good dissimilarity measure for comparing data samples lying on a manifold. In this paper, we propose two algorithms for determining a local subset of training samples from which a good local model can be computed for reconstructing a given input test sample, where we consider the underlying geometry of the data. The first algorithm, called adaptive geometry-driven nearest neighbor search (AGNN), is an adaptive scheme, which can be seen as an out-of-sample extension of the replicator graph clustering method for local model learning. The second method, called geometry-driven overlapping clusters (GOCs), is a less complex nonadaptive alternative for training subset selection. The proposed AGNN and GOC methods are evaluated in image superresolution and shown to outperform spectral clustering, soft clustering, and geodesic distance-based subset selection in most settings. © 1992-2012 IEEE.


Mazivila S.,Federal University of Uberlandia | De Santana F.B.,Federal University of Uberlandia | Mitsutake H.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | And 3 more authors.
Fuel | Year: 2015

The objective of this work was to differentiate the type of biodiesel in mixtures with diesel in the ratio of 5% (v/v) in relation to the type of oil and alcohol used in the production process of the respective biofuel. For this purpose, the mid-infrared (MIR) spectroscopy combined with supervised chemometric tool, partial least squares discriminant analysis (PLS-DA) was used. The distinction between six types of biodiesel/diesel blend was evaluated, in other words, Soybean Oil Ethyl Esters (SOEE), Jatropha Oil Methyl Esters (JOME), Used Frying Oil Methyl Esters (UFOME), Soybean Oil Methyl Esters (SOME), Used Frying Oil Ethyl Esters (UFOEE) and Jatropha Oil Ethyl Esters (JOEE). The PLS-DA models, showed excellent efficiency in the classification for each type of biodiesel, both in relation to type of oil as well as alcohol used in its production, with high levels of sensitivity and specificity for all classes, remitting to 100% correct classification. Therefore, the use of this methodology is a viable alternative to quality control of biodiesel because it allows identify the type of biodiesel present in biodiesel/diesel blend. © 2014 Elsevier Ltd. All rights reserved.


Gontijo L.C.,Federal University of Uberlandia | Gontijo L.C.,Goiano Federal Institute of Education | Guimaraes E.,Federal University of Uberlandia | Mitsutake H.,Federal University of Uberlandia | And 3 more authors.
Fuel | Year: 2014

This study evaluated the use of partial-least-squares (PLS) regression models to quantify soybean biodiesels in diesel blends. The study was carried out by taking into account the entire mid-infrared spectral range and according to the ASTM E1655 standard. The PLS models provided low root-mean-squared errors of prediction (RMSEP) of 0.0792% (v/v) and 0.1050% (v/v) for the models containing methyl and ethyl soybean biodiesels, respectively. In addition, an excellent correlation was observed in the prediction set (R = 0.9999), and no systematic errors were present according to the ASTM E1655 standard. When the models were compared against the requirements of the ABNT NBR 15568 standard, both models exhibited adequate accuracy both the concentration ranges from 0% to 8% and 8 to 30% (v/v). Therefore, the proposed models for the entire spectral region allow the determination of both methyl and ethyl soybean biodiesels in diesel using only the concentration range between 1.00% and 30.00% (v/v). © 2013 Elsevier Ltd. All rights reserved.


De Souza L.M.,Federal University of Uberlandia | Mitsutake H.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | Gontijo L.C.,Goiano Federal Institute of Education | Borges Neto W.,Federal University of Uberlandia
Fuel | Year: 2014

Diesel, inside all energy matrixes, is considered a valuable resource. Brazilian S-10 diesel can be subject to adulteration by addition of residual automotive lubricant (RAL), to increase the profit margin for distributors and fuel service stations. The use of mid-infrared (MIR) spectroscopy in combination with multivariate calibration methods as a quantitative analytical method can provide satisfactory results in accordance with those found using conventional methods. We report the use of MIR spectroscopy and multivariate calibration using the partial least squares method (PLS) for the quantitative analysis of RAL as an adulterant in Brazilian S-10 diesel. Our results displayed good correlation between the reference values and those calculated using the PLS model with low error values. The model was built in accordance with standard method ASTM E1655-05 and validations of the figures of merit were estimated. © 2014 Elsevier Ltd. All rights reserved.


Guimaraes E.,Federal University of Uberlandia | Mitsutake H.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | Gontijo L.C.,Goiano Federal Institute of Education | And 3 more authors.
JAOCS, Journal of the American Oil Chemists' Society | Year: 2015

Abstract This work quantifies the adulteration of ethyl and methyl soybean biodiesels/diesel (B5) blended with soybean oil using mid-infrared spectroscopy associated with multivariate calibration. The models constructed by the method of partial least squares (PLS) presented low values of root-mean-square error of prediction 0.22 % (w/w) and 0.26 % (w/w), respectively, for models containing ethyl and methyl soybean biodiesel. Along with the parameters of error, accuracy was evaluated by the use of an elliptical joint confidence region (EJCR). The EJCR for the both PLS models showed there was no significant difference between the prepared concentration values and PLS predicted concentration values, and that there was no evidence of bias within the 95 % confidence level. The PLS models showed excellent correlation in the prediction set (R = 0.999) and did not present systematic errors according to the ASTM E1655 standard. Therefore, the models presented excellent performance in quantifying soybean oil as an adulterant in B5 blends, in concentrations within the range 1.00-30.00 % (w/w). The proposed methodology showed itself to be efficient for quality control of B5 contaminated with vegetable oil. © 2015 AOCS.


Mazivila S.J.,Federal University of Uberlandia | Mitsutake H.,Federal University of Uberlandia | De Santana F.B.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | And 3 more authors.
Journal of the Brazilian Chemical Society | Year: 2015

This work aimed at employing partial least square discriminant analysis (PLS2-DA), allied to mid-infrared (MIR) spectroscopy as an analytical method for simultaneous classification of biodiesels from different oils (soybean and used frying oil) and routes (methylic and ethylic). The evaluation of the model was verified through values of sensitivity and specificity for each parameter, in the interest class. PLS2-DA model showed 100% correct classification in the discrimination of types of biodiesels. Therefore, the proposed methodology is fast, because it allows simultaneous classification of different types of biodiesels. Consequently, it can be used in quality control of this type of biofuel. © 2015 Sociedade Brasileira de Química.


De Souza L.M.,Federal University of Uberlandia | De Santana F.B.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | Gontijo L.C.,Goiano Federal Institute of Education | And 2 more authors.
Food Chemistry | Year: 2015

This paper proposes a new method for the quantitative analysis of soybean oil (SO) and sunflower oil (SFO) as adulterants in extra virgin flaxseed oil (EFO) by applying Mid Infrared Spectroscopy (MIR) associated with chemometric technique of Partial Least Squares (PLS). The PLS models were built in accordance with standard method ASTM E1655-05 and these showed good correlation between the reference values and those calculated using the PLS models with low error values, with R = 0.998 for SFO and R = 0.999 for SO in EFO. These models were validated analytically in accordance with Brazilian and international guidelines through the estimate of figures of merit parameters, thus showing an effective and feasible method to control the quality of extra virgin flaxseed oil. © 2015 Elsevier Ltd.


Guimaraes E.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | Gontijo L.C.,Goiano Federal Institute of Education | Mitsutake H.,Federal University of Uberlandia | And 3 more authors.
Industrial and Engineering Chemistry Research | Year: 2014

In this work, we developed a method to quantify ethanol in ethyl soybean biodiesel and ethyl biodiesel from waste frying oil using mid-infrared spectroscopy in association with multivariate calibration by the partial least-squares (PLS) method. The obtained models are efficient in ethanol determination in concentrations ranging from 0.14% to 1.00% (w/w). In both PLS models low values of the root-mean-square error prediction (0.02%, w/w) and excellent correlation between measured and predicted values in the prediction set (R > 0.99) were observed, and there were no systematic errors according to the ASTM E1655 standard. The methods were validated according to international and national guidelines by the estimate of figures of merit, such as accuracy, linearity, sensitivity, selectivity, analytical sensitivity, and limits of detection and quantification. Considering these good results, the proposed method can be used for biofuel quality control in a fast, simple, and nondestructive way. © 2014 American Chemical Society.


De Santana F.B.,Federal University of Uberlandia | Gontijo L.C.,Federal University of Uberlandia | Gontijo L.C.,Goiano Federal Institute of Education | Mitsutake H.,Federal University of Uberlandia | And 3 more authors.
Food Chemistry | Year: 2016

Rosehip oil (Rosa eglanteria L.) is an important oil in the food, pharmaceutical and cosmetic industries. However, due to its high added value, it is liable to adulteration with other cheaper or lower quality oils. With this perspective, this work provides a new simple, fast and accurate methodology using mid-infrared (MIR) spectroscopy and partial least squares discriminant analysis (PLS-DA) as a means to discriminate authentic rosehip oil from adulterated rosehip oil containing soybean, corn and sunflower oils in different proportions. The model showed excellent sensitivity and specificity with 100% correct classification. Therefore, the developed methodology is a viable alternative for use in the laboratory and industry for standard quality analysis of rosehip oil since it is fast, accurate and non-destructive. © 2016 Elsevier Ltd. All rights reserved.

Loading Goiano Federal Institute of Education collaborators
Loading Goiano Federal Institute of Education collaborators