Factorial Discriminant Analysis Research Articles

Variable selection and training set design for particle classification using a linear and a non-linear classifier

While particulate products are often characterized by their median diameter or the width of the particle size distribution, information is rarely given about the agglomeration degree of the product. To obtain this information, a tool combining image analysis and discriminant factorial analysis (DFA) was introduced in previous works. The accuracy of that method depended on the number of image descriptors selected, i.e. measurements describing each particle: few image descriptors resulted in rather poor classification while too many lead to an overfitting of the data. The aim of this study is twofold: First, we want to compare the classification accuracy of artificial neural networks (ANN) and DFA which, contrary to ANN, forms linear classifiers. Second, we want to provide an easy-to-implement procedure for generating particle classifiers. We used a qualitative measure called Proportional Similarity to test whether a subset selection of image descriptors was necessary to avoid an overfitting. The influence of the training set size was investigated as well as the transferability of the classifier on data obtained under different experimental conditions. The chemical systems used were l-alanine/water and adipic acid/water and the classes considered were single crystals, agglomerates, and gas bubbles. The results show that an ANN classifier provides higher accuracy and is more effective when only few image descriptors are available while DFA is simpler to create. Moreover, we show good transferability of classifiers trained on data of different experimental conditions. Based on our results, we provide guidelines for classification of particulate systems.

Hyperspectral imaging for rapid evaluation and visualization of quality deterioration index of vacuum packaged dry-cured sausages

A rapid and nondestructive method to evaluate and visualize quality of vacuum packaged dry-cured sausages was developed using hyperspectral imaging (HSI) spectroscopy. Changes in physicochemical, microbiological and sensory attributes of the packaged dry-cured sausages were monitored during storage at 20°C. Discriminant factor analysis (DFA) could classify samples into different groups as a function of storage time. The quality deterioration index (QDI) of the packaged dry-cured sausages was signified and assigned to the individual DFA classified groups. HSI spectroscopy was used to collect both spectral data in the wavelength range of 380–1000nm and spatial data of 68×68 pixels. Partial least squares (PLS) regression model was constructed to establish the relationship between the HSI hypercube data and the QDI. Distribution maps of the QDI were then constructed to visualize the quality attributes of the sausages. The color distribution map of the QDI enabled the identification of different qualities of the vacuum packaged sausages as a function of storage time. The results demonstrated that the HSI coupled with chemometric and image processing techniques can be used as a rapid and nondestructive method to evaluate quality of the sausages.

Ageing status characterization of Chinese spirit using scent characteristics combined with chemometric analysis

Abstract Chinese kaoliang spirit (CTKS) is a popular liquor within the Yue nation with a history dating back 2500 years. However, fraudulent practices have occurred during its commercialization, including the production and sale of adulterated spirits that have been aged differently. Here, a gas–chromatography–flash electronic nose technique was combined with chemometric analysis to develop a rapid method for the discrimination of CTKS of various ages. Models for discrimination were developed using principal component analysis and discriminant factor analysis. We also investigated the volatile organic compounds in CTKS to determine differences between samples of different ages. The electronic nose technique combined with chemometric methods may be used for fingerprinting to authenticate CTKS and protect the quality and prestige of this product.

Global analysis of H3K27me3 as an epigenetic marker in prostate cancer progression

BackgroundH3K27me3 histone marks shape the inhibition of gene transcription. In prostate cancer, the deregulation of H3K27me3 marks might play a role in prostate tumor progression.MethodsWe investigated genome-wide H3K27me3 histone methylation profile using chromatin immunoprecipitation (ChIP) and 2X400K promoter microarrays to identify differentially-enriched regions in biopsy samples from prostate cancer patients. H3K27me3 marks were assessed in 34 prostate tumors: 11 with Gleason score > 7 (GS > 7), 10 with Gleason score ≤ 7 (GS ≤ 7), and 13 morphologically normal prostate samples.ResultsHere, H3K27me3 profiling identified an average of 386 enriched-genes on promoter regions in healthy control group versus 545 genes in GS ≤ 7 and 748 genes in GS > 7 group. We then ran a factorial discriminant analysis (FDA) and compared the enriched genes in prostate-tumor biopsies and normal biopsies using ANOVA to identify significantly differentially-enriched genes. The analysis identified ALG5, EXOSC8, CBX1, GRID2, GRIN3B, ING3, MYO1D, NPHP3-AS1, MSH6, FBXO11, SND1, SPATS2, TENM4 and TRA2A genes. These genes are possibly associated with prostate cancer. Notably, the H3K27me3 histone mark emerged as a novel regulatory mechanism in poor-prognosis prostate cancer.ConclusionsOur findings point to epigenetic mark H3K27me3 as an important event in prostate carcinogenesis and progression. The results reported here provide new molecular insights into the pathogenesis of prostate cancer.

Seven decades of change in forest structure and composition in Pinus resinosa forests in northern Minnesota, USA: Comparing managed and unmanaged conditions

An understanding of long-term patterns of forest structural and compositional development is critical for anticipating management outcomes and developing appropriate silvicultural strategies for restoring complex forest conditions. In most cases, this information comes from stand-level assessments; however, the impacts and outcomes of management and other disturbances on forest development occur over multiple spatial scales across a landscape. We compared historical (1941) and contemporary (2012–2014) forest structure and composition on 301 plots distributed across managed and unmanaged, late seral red pine (Pinus resinosa)-dominated forests in a 1230ha landscape in north-central Minnesota, USA. Discriminate factor analysis was used to determine which compositional and structural attributes best described the forest conditions between two sampling periods (1941, 2012–2014) and management histories (managed and unmanaged). Plot basal area, average diameter of live trees, richness of tree size classes, and the basal area of standing deadwood were the four most important variables in discriminating between the managed and unmanaged plots in 1941 and 2013. In some cases, structural conditions between managed and unmanaged plots converged, including contemporary BA, trees per hectare, size inequality, and structural complexity indices. In contrast, several attributes, including standing deadwood basal area and percent hardwood basal area, were significantly greater in unmanaged plots after 72years and highlight the lasting influence of land use on these structural and compositional conditions. The broad ranges of structural and compositional conditions observed across the landscape highlight the importance of having spatially varying desired future conditions across managed stands to approximate this range in live and dead-tree attributes in unmanaged forests. In addition, the lower basal area of standing dead trees documented in this and other comparisons of unmanaged and managed P. resinosa systems argues for an increased emphasis on deliberate deadwood and live-tree retention to recruit these features if the restoration of late-successional forest conditions is included as a management objective.

Patterns of genetic structure and phenotypic diversity in sorghum landraces in relation to farmers’ management in Burkina Faso

Sorghum [Sorghum bicolor (L.) Moench] is an important food crop in Burkina Faso. Farmers grow various sorghum landraces, distinguished by their morphological and phenological characteristics. This study aimed addressing how this varietal diversity is managed by farmers, and how their management shape the genetic structures of local sorghum. Diversity levels in ten local sorghum guinea varieties (25 panicles per variety) collected from different farms in three agro-ecological regions across Burkina Faso were assessed using nine agro-morphological traits and twelve SSR markers. The importance of phenotypic intra-varietal diversity in early, intermediate and late maturing varieties varied in relation to the observed trait. The genetic analysis specified that the highest diversity rates were in the late maturing varieties (3.3 for allelic richness and 0.47 for gene diversity). Factorial Discriminant Analysis on quantitative traits and cluster analysis based on SSR markers showed that early maturing varieties were discriminated from late and intermediate ones, confirming farmers' classification of maturity groups (MG). Multivariate QST and FST values revealed two subgroups of quantitative traits that underwent either stabilizing or divergent selection among MG scale. These results underline the role of farmer practices in phenotypic and genetic evolution of sorghum. This concept should be well considered in sorghum breeding programs. (Resume d'auteur)

Efficiency of Rosemary and Basil Essential Oils on the Shelf-Life Extension of Atlantic Mackerel (Scomber scombrus) Fillets Stored at 2°C.

The objective of the present study was to evaluate the impact of rosemary and basil essential oils (EOs) on the quality of Atlantic mackerel fillets stored at 2°C up to 15 days. Atlantic mackerel (Scomber scombrus) fillets were periodically evaluated to assess their textural, color, physicochemical, and spectral characteristics. The results indicated that rosemary and basil treatments were effective for inhibiting the formation of total volatile basic nitrogen (TVB-N) and lipid oxidation products during storage. Based on TVB-N values, the shelf life of Atlantic mackerel fillets treated with rosemary and basil EOs was extended by 2 and 5 days, respectively, compared to the control group. Similar results were obtained with thiobarbituric acid-reactive substance analysis, which demonstrated an extended shelf life of Atlantic mackerel immersed with rosemary and basil EOs of 2 and 3 days, respectively, compared to the control group. The factorial discriminant analysis applied on the concatenated first five principal components corresponding to the physicochemical, textural, color, and fluorescence measurements allowed clear discrimination of the three groups, because a correct classification rate of 93.3% was obtained. Therefore, treatment with basil and rosemary EOs, as natural biopreservative compounds, could present a high-potential application in the seafood industry.

Discrimination of producing areas of Astragalus membranaceus using electronic nose and UHPLC-PDA combined with chemometrics

Electronic nose (E-nose) and ultra-high-performance liquid chromatography (UHPLC) were simultaneously developed for quality evaluation and diversity research of A. mongholicus. Instrumental data were subjected to principal component analysis (PCA) and discriminant factor analysis (DFA) to rapidly distinguish A. mongholicus derived from different main producing regions in China. E-nose results showed that A. mongholicus that originated from main producing areas (Gansu, Shanxi, and Inner Mongolia in China) could be accurately classified into three groups through PCA and DFA. The response values of samples from Inner Mongolia and Shanxi were similar, and better than those of the samples from Gansu. Identification by the E-nose was verified through UHPLC analysis, and the content of the main ingredient was consistent with the response value of A. mongholicus. In conclusion, E-nose and UHPLC combined with chemometrics can be used as reliable techniques to distinguish the geographic origins of A. mongholicus effectively, and evaluate the quality of traditional Chinese herbs.

Differentiation of chill‐stored and frozen pork necks using electronic nose with ultra‐fast gas chromatography

AbstractThe purpose of this study was to determine the possibility of differentiating chill‐stored and frozen pork necks with a technique of electronic nose (E‐nose) based on ultra‐fast gas chromatography (UFGC). Tests were carried out on pork necks from six pigs obtained from a breeding farm with Pork Quality System (PQS) implemented. The volatile organic compounds (VOC) of the meat samples with 1, 4, and 7 days of chill storage and samples with 5 months of frozen storage were investigated. Principal components analysis (PCA) showed differences in the volatile organic compounds profiles. Discriminant factor analysis (DFA) model was developed to identify the freshness of the meat samples. Results showed that using E‐nose Heracles II allowed effective recognition of chilled and frozen pork meats. Significant differences were found between VOC profiles. In practice, this technique could be applied to the meat industry for a rapid, cheap, and reliable assessment of meat freshness.Practical applicationsInvestigation of methods for differentiation of chill‐stored and frozen pork necks is of significance for quality and safety maintenance. The current study developed a technique of electronic nose (E‐nose) based on ultra‐fast gas chromatography (UFGC) for recognizing the freshness difference of these meat products. Results showed that the technique allowed for effective recognition of chilled and frozen pork meats. In practice, this technique could be applied to the meat industry for a rapid, cheap, and reliable assessment of meat freshness.

3D Face Recognition Based on Symbolic FDA Using SVM Classifier with Similarity and Dissimilarity Distance Measure

Human face images are the basis not only for person recognition, but for also identifying other attributes like gender, age, ethnicity, and emotional states of a person. Therefore, face is an important biometric identifier in the law enforcement and human–computer interaction (HCI) systems. The 3D human face recognition is emerging as a significant biometric technology. Research interest into 3D face recognition has increased during recent years due to availability of improved 3D acquisition devices and processing algorithms. A 3D face image is represented by 3D meshes or range images which contain depth information. In this paper, the objective is to propose a new 3D face recognition method based on radon transform and symbolic factorial discriminant analysis using KNN and SVM classifier with similarity and dissimilarity measures, which are applied on 3D facial range images. The experimentation is done using three publicly available databases, namely, Bhosphorus, Texas and CASIA 3D face database. The experimental results demonstrate the effectiveness of the proposed method.

Front face fluorescence spectroscopy enables rapid differentiation of fresh and frozen-thawed sea bass (Dicentrarchus labrax) fillets

The aim of this study was to evaluate the potential of using front face fluorescence spectroscopy (FFFS) to discriminate fresh refrigerated (up to 13 days at 4 °C) sea bass (Dicentrarchus labrax) fillets (F group) from: i) frozen-thawed (3 months at −18 °C), then refrigerated up to 9 days fillets at 4 °C (F-T R group), and ii) refrigerated (up to 9 days at 4 °C), then frozen-thawed (3 months at −18 °C) fillets (R F-T group). The evolution of emission fluorescence spectra after excitation set at 290, 340, and 380 nm, and excitation spectra after emission set at 410 nm were acquired throughout storage. Some traditional measurements based on color, textural, and chemical parameters were performed on the same fillets. Canonical correlation analysis (CCA), applied to the traditional and fluorescence data sets, showed high correlations between the two data sets, since squared canonical coefficients were found higher than 0.8 for canonical variates 1 and 2. This was confirmed by the factorial discriminant analysis (FDA) applied to the concatenated emission spectra recorded after excitation sets at 340 and 380 nm since 72 out of 78 samples were correctly classified. Finally, a multi-block analysis method based on common components and specific weights analysis (CCSWA) was applied to the whole data sets. A reliable discrimination between fresh from frozen/thawed fish was found suggesting that FFFS could be used as a promising tool in routine inspection for differentiating between fresh and frozen-thawed fish.

Voice Habits and Behaviors: Voice Care Among Flamenco Singers

The purpose of this study is to analyze the vocal behavior of flamenco singers, as compared with classical music singers, to establish a differential vocal profile of voice habits and behaviors in flamenco music. Bibliographic review was conducted, and the Singer's Vocal Habits Questionnaire, an experimental tool designed by the authors to gather data regarding hygiene behavior, drinking and smoking habits, type of practice, voice care, and symptomatology perceived in both the singing and the speaking voice, was administered. We interviewed 94 singers, divided into two groups: the flamenco experimental group (FEG, n = 48) and the classical control group (CCG, n = 46). Frequency analysis, a Likert scale, and discriminant and exploratory factor analysis were used to obtain a differential profile for each group. The FEG scored higher than the CCG in speaking voice symptomatology. The FEG scored significantly higher than the CCG in use of "inadequate vocal technique" when singing. Regarding voice habits, the FEG scored higher in "lack of practice and warm-up" and "environmental habits." A total of 92.6% of the subjects classified themselves correctly in each group. The Singer's Vocal Habits Questionnaire has proven effective in differentiating flamenco and classical singers. Flamenco singers are exposed to numerous vocal risk factors that make them more prone to vocal fatigue, mucosa dehydration, phonotrauma, and muscle stiffness than classical singers. Further research is needed in voice training in flamenco music, as a means to strengthen the voice and enable it to meet the requirements of this musical genre.

Effect of high pressure treatment on the aging characteristics of Chinese liquor as evaluated by electronic nose and chemical analysis.

Several high pressure (HP) treatments (100–400 MPa; 15 and 30 min) were applied to Chinese “Junchang” liquor, and aging characteristics of the liquor were evaluated. Results from the principal component analysis and the discriminant factor analysis of E-Nose demonstrated that HP treatment at 300 and 400 MPa resulted in significant (p < 0.05) changes in aroma components of the liquor. An increase in total ester content and a decrease in total acid content were observed for all treated samples (p < 0.05), which was verified by gas chromatography analysis. In addition, a slight decrease in alcohol content was found for HP treatment at 400 MPa for 30 min. These changes and trends were in accordance with the natural aging process of Chinese liquor. However, HP treatment caused a slight increase in solid content, which might be somewhat undesirable. Sensory evaluation results confirmed that favorable changes in color and flavor of Chinese liquor were induced by HP treatment; however, overall gaps still existed between the quality of treated and six-year aged samples. HP treatment demonstrated a potential to accelerate the natural aging process for Chinese liquor, but long term studies may be needed further to realize the full potential.

Discriminant Analysis of Factors Affecting Traffic Accident Severity During Daytime and Nighttime

PURPOSES : Low visibility caused by dark surroundings at nighttime affects the likelihood of accidents, and various efforts, such as installing road safety facilities, have been made to reduce accidents at night. Despite these efforts, the nighttime severity index (SI) in Korea was higher than the daytime SI during 2011-2014. This study determined the factors affecting daytime and nighttime accident severity through a discriminant analysis. METHODS : Discriminant analysis. RESULTS : First, drowsiness, lack of attention, and lighting facilities affected both daytime and nighttime accident severity. Accidents were found to be caused by a low ability to recognize the driving conditions and a low obstacle avoidance capability. Second, road conditions and speeding affected only the daytime accident severity. Third, failure to maintain a safe distance significantly affected daytime accident severity and nonsignificantly affected nighttime accident severity. The majority of such accidents were caused by rear-end collisions of vehicles driving in the same direction; given the low relative speed difference in such cases, the shock imparted by the accidents was minimal. CONCLUSIONS : Accidents caused by a failure to maintain a safe distance has lower severity than do accidents caused by other factors.

A multivariate approach for the study of environmental drivers of wine economic structure

Vitivinicultural ⿿terroir⿿ is a concept referring to an area in which the collective knowledge of the interactions between environment and vitivinicultural practices develops, providing distinctive characteristics to the products. The effect of the environmental components over the terroir wines has been already widely demonstrated, but their possible effect on the overall system of production and organization of the viticultural farms and wine industry, called ⿿wine economic structure⿿ in this study (WES), is still unknown. Thus the research hypothesis of this work was that there are environmental drivers influencing WES.The investigation was carried out at the national scale, taking Italy as study area, the biggest wine producer of the world. We used viticultural territories singled out at 1:500,000 reference scale, called viticultural Macroareas, which grouped Denomination of Origin areas with respect of geographic proximity, environmental features, viticultural affinity, and wine tradition. The characterization of WES was based on the official data reported in the wine production declarations related to the year 2008. Statistics were taken into account about general quantitative variables of wine farms, presence of associative forms, degree of vertical integration of wineries, quality orientation of wine producers, and acreage of vineyard. Environmental variables related to climate, soil, and vegetation vigour were selected for their direct influence on the vine growing. A second set of variables was chosen to express the effect of land topography on viticultural management. A third one was intended to discover the possible relationships between WESs and land qualities, such as the indexes of sensitivity to desertification, the soil resistance to water erosion, and land vulnerability. A multivariate approach was carried out to elaborate the many studied variables and their interactions. Nonparametric statistics were also used, to overcome the necessity of normal distributions. A PCA was carried out separately for the environmental and economic data, to reduce the dimension of datasets. The new economic and environmental synthetic descriptors were submitted to three multivariate analyses: (i) the correlation between economic and environmental descriptors, through the non-parametric Spearman test; (ii) a cluster analysis to group the Macroareas in few homogeneous WESs; (iii) a discriminant analysis of economic clusters and environmental factors, to highlight the environmental drivers of WESs. The results highlighted that in some Macroareas there is a direct correlation between high quality wine production and presence of low available soil water, due to both soil and climatic limitations. In other Macroareas, the limited quantity and quality of wine yield showed a significant relationship with the presence of rough morphologies and high environmental variability. Similarly, heterogeneous soil and morphological conditions fitted well with both the absence of well-defined economic structures, and a relatively lower quality production. The cluster analysis identified six main WESs for the whole country. Climatic, pedoclimatic, morphological mean conditions and topographic heterogeneity of Macroareas had the most important discriminant power over the clusters. The result of this multivariate analyses proved that the different WESs of a country are significantly influenced by specific landscape characteristics. Therefore, landscape care has a strategic role also on the development of the wine industry.

Caracterisation Agro-Morphologique Des Accessions D’arachide (Arachis Hypogaea L.) Pour La Teneur En Huile

In Niger, groundnut [Arachis hypogaea (L.)] is a leguminous crop cultivated in large area. Its seeds are used by oil industries. Groundnut oil is well appreciated by consumers notably in rural areas. The aim of this study was to select the accessions and taxa with high oil content and identify the oil content related traits. Thus, 20 accessions (9 local and 11 improved) belonging to 5 taxa types (FST, VUL, PRU, HYR, and HYB) were planted in randomized block design with three replications. Fifteen traits were investigated in field while oil content was quantified in laboratory. An analysis of variance was performed on all measured traits. Discriminant factorial analysis (AFD) was performed on taxon types, correlation between oil content and other traits was determined. Analysis revealed high and significant correlation between oil content and seed size. In the study, groundnut accessions ICG 8352 and ICG 9991 (oil content 52.8%) and taxon Fastigiata (oil content 48.52%) recorded highest oil content. Seed size and taxon type in groundnut could be criteria to select genotypes for oil industries.

English

A hundred and seventy-one oleaginous watermelon accessions either collected from different countries or obtained from gene banks were evaluated and compared based on 11 quantitative morphological traits. Principal component analysis on 11 traits revealed 81.19% of the total variability and pointed out variations among accessions, mainly on the basis of fruit size and weight. The dendrogram and factorial discriminant analysis clustered the accessions in four groups. The multivariate analysis of variance showed a significant difference between the four groups and accessions of the group 2 had higher agronomic performances. The confusion matrix gave the details about accessions assignment and pointed out that breeders must have recourse to several multivariate analyses to have better agromorphological classification of accessions. The traits related to fruit size and weight were the most accession distinctive. Consequently, fruit morphological traits could be used for accession identification during collecting missions. Key words: Citrullus mucosospermus, morphological characterization, multivariate analysis, oleaginous cucurbit.

Discrimination of edible fungi varieties and evaluation of their umami intensities by using an electronic tongue method

SummaryThe flavour and taste of edible fungi are highly associated with the presence of several umami substances. In this study, the different edible fungi are discriminated by an electronic tongue method. The response data are analysed by the PCA (principal component analysis), DFA (discriminant factor analysis) and PLS (partial least square) method. The results show that the two data analysis methods (PCA and DFA) could successfully distinguish the variety of edible fungi. Moreover, the umami intensities of the edible fungi solutions are evaluated and compared by an electronic tongue analysis with calibration by the responses of different monosodium glutamate (MSG) concentrations. The signals are also used to build up PLS models to estimate and predict the umami intensities of the edible fungi. The results of this study indicate that the electronic tongue has a great potential in qualitative and quantitative analysis of the sensory quality of edible fungi.

Development of a new fertility prediction model for stallion semen, including flow cytometry

Several laboratories routinely use flow cytometry to evaluate stallion semen quality. However, objective and practical tools for the on-field interpretation of data concerning fertilizing potential are scarce. A panel of nine tests, evaluating a large number of compartments or functions of the spermatozoa: motility, morphology, viability, mitochondrial activity, oxidation level, acrosome integrity, DNA integrity, “organization” of the plasma membrane, and hypoosmotic resistance, was applied to a population of 43 stallions, 33 of which showing widely differing fertilities (19%–84% pregnancy rate per cycle [PRC]). Analyses were performed either within 2 hours after semen collection or after 24-hour storage at 4 °C in INRA96 extender, on three to six ejaculates for each stallion. The aim was to provide data on the distribution of values among said population, showing within-stallion and between-stallion variability, and to determine whether appropriate combinations of tests could evaluate the fertilizing potential of each stallion. Within-stallion repeatability, defined as intrastallion correlation (r = between-stallion variance/total variance) ranged between 0.29 and 0.84 for “conventional” variables (viability, morphology, and motility), and between 0.15 and 0.81 for “cytometric” variables. Those data suggested that analyzing six ejaculates would be adequate to characterize a stallion. For most variables, except those related to DNA integrity and some motility variables, results differed significantly between immediately performed analyses and analyses performed after 24 hours at 4 °C. Two “best-fit” combinations of variables were determined. Factorial discriminant analysis using a first combination of seven variables, including the polarization of mitochondria, acrosome integrity, DNA integrity, and hypoosmotic resistance, permitted exact determination of the fertility group for each stallion: fertile, that is, PRC higher than 55%; intermediate, that is, 45% < PRC less than 55%; or subfertile, that is, PRC less than 45%. Linear regression using another combination of 20 variables, including motility, viability, oxidation level, acrosome integrity, DNA integrity, and hypoosmotic resistance, accounted for 94.2% of the variability regarding fertility and was used to calculate a prediction of the PRC with a mean standard deviation of 3.1. The difference between the observed fertility and the calculated value ranged from −4.2 to 5.0. In conclusion, this study enabled to determine a new protocol for the evaluation of stallion semen, combining microscopical observation, computer-assisted motility analysis and flow cytometry, and providing a high level of fertility prediction.

Discrimination and Identification of Vegetable Oil Based on Voltammetric Electronic Tongue

The study presented the application of a voltammetric electronic tongue to discriminate and identify vegetable oil. Concretely, it aimed to research the discrimination of different oil and the prediction of unknown oil. Seven oil samples from different varieties and geographical origins were measured by a voltammetric platinum electrode as the sensing part. The electrochemical response current signals of samples which were the original data information were obtained with cyclic voltammetric measurement. Principal Component Analysis (PCA) and Cluster Analysis (CA) algorithms were used as the modeling tools to discriminate different vegetable oil respectively. Discriminant Factorial Analysis (DFA) and RBF Neural Network (RBFNN) were used as the prediction models for unknown oil. Fast Fourier Transform (FFT) and Discrete Wavelet Transform (DWT) were applied as feature extraction method for data input set of the prediction models. Different combinations of prediction strategies with feature extraction methods were compared. It was found the samples with different varieties or origins were clearly discriminated with using PCA and CA. The best prediction results were obtained with a 90.48% of identification accuracy by employing FFT-RBFNN. The implementation of this study suggests the electronic tongue may be a useful tool for oil quality evaluation and control.