Nathalie Dupuy

Automated Principal Component-Based Orthogonal Signal Correction Applied to Fused Near Infrared−Mid-Infrared Spectra of French Olive Oils

An approach for automating the determination of the number of components in orthogonal signal correction (OSC) has been devised. In addition, a novel principal component OSC (PC-OSC) is reported that builds softer models for removing background from signals and is much faster than the partial least-squares (PLS) based OSC algorithm. These signal correction methods were evaluated by classifying fused near- and mid-infrared spectra of French olive oils by geographic origin. Two classification methods, partial least-squares-discriminant analysis (PLS-DA) and a fuzzy rule-building expert system (FuRES), were used to evaluate the signal correction of the fused vibrational spectra from the olive oils. The number of components was determined by using bootstrap Latin partitions (BLPs) in the signal correction routine and maximizing the average projected difference resolution (PDR). The same approach was used to select the number of latent variables in the PLS-DA evaluation and perfect classification was obtained. Biased PLS-DA models were also evaluated that optimized the number of latent variables to yield the minimum prediction error. Fuzzy or soft classification systems benefit from background removal. The FuRES prediction results did not differ significantly from the results that were obtained using either the unbiased or biased PLS-DA methods, but was an order of magnitude faster in the evaluations when a sufficient number of PC-OSC components were selected. The importance of bootstrapping was demonstrated for the automated OSC and PC-OSC methods. In addition, the PLS-DA algorithms were also automated using BLPs and proved effective.

Assessing petroleum oils biodegradation by chemometric analysis of spectroscopic data.

This study was conducted to classify petroleum oils in terms of their biodegradation stage by using spectroscopic analysis associated to chemometric treatments. Principal Component Analysis (PCA) has been applied on infrared and UV fluorescence spectra of Brazilian and Pyrenean oils. For Brazil samples, the method allowed to distinguish the biodegraded oils from the non-affected ones. Pyrenean sampling including oils at different levels of biodegradation has been chosen to follow their alteration rate. PCA loadings have shown spectral regions which have differentiated oils after biodegradation whereas Simple-to-use Interactive Self-Modelling Mixture Analysis (SIMPLISMA) has permitted to obtain a repartition in terms of components families (saturated, aromatic and polar ones) characterizing chemical composition of oils at different biodegradation degrees. Results are in good agreement with conclusions of usual hydrocarbon biomarker analysis.

Fire impact on forest soils evaluated using near-infrared spectroscopy and multivariate calibration

The assessment of physico-chemical properties in forest soils affected by fires was evaluated using near infrared reflectance (NIR) spectroscopy coupled with chemometric methods. In order to describe the soil properties, measurements were taken of the total organic carbon on solid phase, the total nitrogen content, the organic carbon and the specific absorbences at 254 and 280 nm of humic substances, organic carbon in humic and fulvic acids, concentrations of NH(4)(+), Ca(2+), Mg(2+), K(+) and phosphorus in addition to NIR spectra. Then, a fire recurrence index was defined and calculated according to the different fires extents affecting soils. This calculation includes the occurrence of fires as well as the time elapsed since the last fire. This study shows that NIR spectroscopy could be considered as a tool for soil monitoring, particularly for the quantitative prediction of the total organic carbon, total nitrogen content, organic carbon in humic substances, concentrations of phosphorus, Mg(2+), Ca(2+) and NH(4)(+) and humic substances UVSA(254). Further validation in this field is necessary however, to try and make successful predictions of K(+), organic carbon in humic and fulvic acids and the humic substances UVSA(280). Moreover, NIR coupled with PLS can also be useful to predict the fire recurrence index in order to determine the spatial variability. Also this method can be used to map more or less burned areas and possibly to apply adequate rehabilitation techniques, like soil litter reconstitution with organic enrichments (industrial composts) or reforestation. Finally, the proposed recurrence index can be considered representative of the state of the soils.

Development of near infrared sensors: detection of influential factors by the AComDim method.

The development of near infrared (NIR) sensors has to go through different steps of testing. Once a prototype is ready to be used, it is necessary to evaluate and optimize the experimental conditions and the data collection, in terms of accuracy, repeatability, reproducibility and speed. This paper studies the effects of controllable experimental factors on the quality of the spectral response, to determine the influence of each instrumental parameter and to improve the predictions obtained from the collected data. The AComDim method, based on the multi-block analysis of ANOVA matrices, was used here to evaluate the impact of experimental factors on the responses from the different sensors tested.

Fatty Acid Compositions of Olive Oils from Six Cultivars from East and South-Western Algeria

The aim of this work is to characterize olive oils from six algerian cultivars (Azeradj, n=4); Blanquette, n=7; Bouricha, n=2; Chemlal, n=5; Limli, n=3; Sigoise, n=1) by determining their fatty acid compositions. The fatty acid composition of oils is determined using gas chromatography of methyl esters obtained by transesterification of triacylglycerols with 2M KOH/MeOH. Fourteen fatty acids and squalene are identified in all the samples. Oleic acid (18:1ω9), palmitic acid (16:0), linoleic acid (18:2ω6) and stearic acid (18:0) are the major fatty acids commonly found in olive oils. Palmitoleic acid (16:1ω7), hypogeic acid (16:1ω9), oleic acid (18:1ω9) and cis-vaccenic acid (18:1ω7) are considered as separate entities by the European regulation, unlike the Codex Alimentarius which identifies them as a single component. Six minor fatty acids namely margaric acid (17:0), margaroleic acid (17:1ω8), arachidic acid (20:0), gondoique acid (20:1ω9), behenic acid (22:0) and lignoceric acid (24:0) are identified. These acids, although minor, are important for the characterization of cultivars. Oils of different cultivars are characterized by different fatty acid compositions. All the values of fatty acid compositions are in compliance with the regulations of the International Olive Oil Council and Codex Alimenta-rius. A radial plot enables the analysis and characterization of each variety as a morphotype by creating a morphogramme . The morphogramme is designed as a radial representation of each fatty acid (n=14) using an Excel ® spreadsheet, and each axis represents the change in the variable with respect to the mean. The morphotypes are real fingerprints of different oil cultivars. Thus, the oils of Azeradj, Blanquette, Chemlal, Limli and Sigoise cultivars have specific morphotype . The morphotype of the Bourricha cultivar is very similar to the Chemlal cultivar. In addition, the Blanquette morphotype is identical to that of the tunisian cultivar Chetoui. This mode of representation is particularly effective for the rapid visual identification of characteristics of olive oils.