Improving the performance of self-organizing maps via growing representations

Self-organizing maps (SOMs) are a type of artificial neural network that through training can produce simplified representations of large, high dimensional data sets. These representations are typically used for visualization, classification, and clustering and have been successfully applied to a variety of problems in the pharmaceutical and bioinformatics domains. SOMs in these domains have generally been restricted to static sets of nodes connected in either a grid or hexagonal connectivity and planar or toroidal topologies. We investigate the impact of connectivity and topology on SOM performance, and experiments were performed on fixed and growing SOMs. Three synthetic and two relevant data sets from the chemistry domain were used for evaluation, and performance was assessed on the basis of topological and quantization errors after equivalent training periods. Although we found that all SOMs were roughly comparable at quantizing a data space, there was wide variation in the ability to capture its underlying structure, and growing SOMs consistently outperformed their static counterparts in regards to topological errors. Additionally, one growing SOM, the Neural Gas, was found to be far more capable of capturing details of a target data space, finding lower dimensional relationships hidden within higher dimensional representations.     

Decision trees in selection of featured determined food quality

The determination of food quality, authenticity and the detection of adulterations are problems of increasing importance in food chemistry. Recently, chemometric classification techniques and pattern recognition analysis methods for wine and other alcoholic beverages have received great attention and have been largely used. Beer is a complex mixture of components: on one hand a volatile fraction, which is responsible for its aroma, and on the other hand, a non-volatile fraction or extract consisting of a great variety of substances with distinct characteristics. The aim of this study was to consider parameters which contribute to beer differentiation according to the quality grade. Chemical (e.g. pH, acidity, dry extract, alcohol content, CO(2) content) and sensory features (e.g. bitter taste, color) were determined in 70 beer samples and used as variables in decision tree techniques. This pattern recognition techniques applied to the dataset were able to extract information useful in obtaining a satisfactory classification of beer samples according to their quality grade. Feature selection procedures indicated which features are the most discriminating for classification.     

The performance of nu-support vector regression on determination of soluble solids content of apple by acousto-optic tunable filter near-infrared spectroscopy

The ν-support vector regression (ν-SVR) was used to construct the calibration model between soluble solids content (SSC) of apples and acousto-optic tunable filter near-infrared (AOTF-NIR) spectra. The performance of ν-SVR was compared with the partial least square regression (PLSR) and the back-propagation artificial neural networks (BP-ANN). The influence of SVR parameters on the predictive ability of model was investigated. The results indicated that the parameter ν had a rather wide optimal area (between 0.35 and 1 for the apple data). Therefore, we could determine the value of ν beforehand and focus on the selection of other SVR parameters. For analyzing SSC of apple, ν-SVR was superior to PLSR and BP-ANN, especially in the case of fewer samples and treating the noise polluted spectra. Proper spectra pretreatment methods, such as scaling, mean center, standard normal variate (SNV) and the wavelength selection methods (stepwise multiple linear regression and genetic algorithm with PLS as its objective function), could improve the quality of ν-SVR model greatly.     

High dimensional model representation constructed by support vector regression. I. Independent variables with known probability distributions

High dimensional model representation (HDMR) is a general set of quantitative model assessment and analysis tools for capturing high dimensional input-output system behavior. In practice, the HDMR component functions are each approximated by an appropriate basis function expansion. This procedure often requires many input-output samples which can restrict the treatment of high dimensional systems. In order to address this problem we introduce svr-based HDMR to efficiently and effectively construct the HDMR expansion by support vector regression (SVR) for a function \(f(\mathbf{x})\). In this paper the results for independent variables sampled over known probability distributions are reported. The theoretical foundation of the new approach relies on the kernel used in SVR itself being an HDMR expansion (referred to as the HDMR kernel ), i.e., an ANOVA kernel whose component kernels are mutually orthogonal and all non-constant component kernels have zero expectation. Several HDMR kernels are constructed as illustrations. While preserving the characteristic properties of HDMR, the svr-based HDMR method enables efficient construction of high dimensional models with satisfactory prediction accuracy from a modest number of samples, which also permits accurate computation of the sensitivity indices. A genetic algorithm is employed to optimally determine all the parameters of the component HDMR kernels and in SVR. The svr-based HDMR introduces a new route to advance HDMR algorithms. Two examples are used to illustrate the capability of the method.     

Ecotoxicity and chemical sediment data classification by the use of self-organising maps

The present paper deals with the presentation in a new interpretation of sediment quality assessment. This original approach studies the relationship between ecotoxicity parameters (acute and chronic toxicity) and chemical components (polluting species like polychlorinated biphenyls (PCBs), pesticides, polycyclic aromatic hydrocarbons (PAH), heavy metals) of lake sediments samples from Turawa Lake, Poland by an application of self-organising maps (SOMs) to the monitoring dataset (59 samples × 44 parameters) in order to obtain visual images of the components distributed at each sampling site when all components are included in the classification and data projection procedure. From the SOMs obtained, it is possible to select groups of similar ecotoxicity (either acute or chronic) and to analyse within each one of them the relationship of the other chemicals to the toxicity determining parameters (EC₅₀ and mortality). Studies have shown, convincingly, that different regions from the Turawa Lake bottom indicate different patterns of ecotoxicity related to various chemical pollutants, such as the “heptachlor-B” pattern, “pesticide and PAH” pattern, “structural” pattern or “PCB congeners” pattern. Thus, an easy way of multivariate analysis of small datasets with ecotoxicity parameters involved becomes possible. Additionally, a distinction between the effects of pollution on acute and chronic toxicity seems reasonable.     

Single-particle detection of cholesterol based on the host-guest recognition induced plasmon resonance energy transfer

Plasmon resonance energy transfer (PRET) occurs between the plasmonic nanoparticles (NPs) and organic dyes forming donor-acceptor pairs, which has great potential in quantitative analytical chemistry because of its excellent sensitivity under dark-field microscopy (DFM). Herein, we introduce supramolecular β-cyclodextrin (β-CD) to design a host-guest recognition plasmonic nano-structure modified gold nanoparticles (GNPs), while GNPs and rhodamine molecule (RB) act as the donor and acceptor, respectively. In the presence of the target cholesterol, due to the stronger binding of cholesterol with β-CD, RB molecules are released, inducing the inhibition of PRET, as well as the increase of the scattering intensity of GNPs. The proposed strategy achieves a linear range from 0.02 µmol/L to 2.0 µmol/L for cholesterol detection, and reaches a limit of detection (LOD) of 6.7 nmol/L. This host-guest recognition strategy can easily integrate receptor-donor pair into one nanoparticle, which simplifies the construction of the PRET platform, and further provides an effective approach for PRET-based analytical applications. Afterwards, the proposed PRET strategy was successfully applied for the detection of cholesterol in serum samples with high sensitivity and specificity. The proposed method provides an effective clinically potential means for the detection of cholesterol and other disease-related biomarkers.     

Pattern recognition in analytical chemistry

Pattern recognition methods are applied in order to classify unknown objects into categories, or to separate objects into categories. Some basic principles and simple methods of pattern recognition are described; typical applications in analytical chemistry are discussed; warnings about improper usage of pattern recognition methods are also emphasized.     

Classification of weathered petroleum oils by multi-way analysis of gas chromatography-mass spectrometry data using PARAFAC2 parallel factor analysis

The application of multi-way parallel factor analysis (PARAFAC2) is described for the classification of different kinds of petroleum oils using GC-MS. Oils were subjected to controlled weathering for 2, 7 and 15 days and PARAFAC2 was applied to the three-way GC-MS data set (MSxGCxsample). The classification patterns visualized in scores plots and it was shown that fitting multi-way PARAFAC2 model to the natural three-way structure of GC-MS data can lead to the successful classification of weathered oils. The shift of chromatographic peaks was tackled using the specific structure of the PARAFAC2 model. A new preprocessing of spectra followed by a novel use of analysis of variance (ANOVA)-least significant difference (LSD) variable selection method were proposed as a supervised pattern recognition tool to improve classification among the highly similar diesel oils. This lead to the identification of diagnostic compounds in the studied diesel oil samples.     

An improved method for the simultaneous analysis of phenolic and steroidal estrogens in water and sediment

This paper describes an improved method for the extraction and analysis of seven endocrine disrupting chemicals with wide-ranging polarities from water and sediments using gas chromatography-tandem mass spectrometry (GC-MS/MS). The analytes were 4-tert-octylphenol, 4-nonylphenol, bisphenol A, estrone, 17β-estradiol, 17α-ethynylestradiol and 16α-hydroxyestrone. The optimised GC-MS/MS method produces increased selectivity and sensitivity compared to GC-MS, with limit of detection ranging from 0.01 to 0.49 ng L⁻¹ in water and from 0.05 to 0.14 ng g⁻¹ in sediment. Extraction from aqueous samples was performed by solid-phase extraction (SPE) and from sediment samples by microwave-assisted extraction (MAE). The improved method for the clean-up of sediment extracts carried out by SPE enhanced EDC recovery (86-102%) while reducing matrix interference and sample drying time. Derivatisation of final sample extracts was achieved using N,O-bis(trimethylsilyl)trifluoroacetamide and pyridine, and their stability was enhanced by reconstituting the derivatised extracts with hexane. The method was validated by spiking experiments which showed good recovery and reproducibility. The method was applied to samples taken from the Medway estuary in Kent, UK, where non-conservative behaviour of EDCs was demonstrated.     

Fisher判别在气相色谱-质谱分析助燃剂燃烧残留物中的应用

为寻找一种用于火场助燃剂燃烧残留物鉴定的更为准确、有效的模式识别方法,对7种常见助燃剂在不同载体上的燃烧残留物样品及未知送检样品进行气相色谱-质谱（GC-MS）分析测试,通过特征组分分析鉴定出未知样品中含有汽油成分。同时运用Fisher判别及PCA（主成分分析）/Fisher判别联用两种判别方法对样本数据进行了分析处理,PCA/Fisher判别联用的结果表明送检样本中含有硝基油漆稀料成分,而仅使用Fisher判别的结果表明送检样本中含有93^#汽油。通过将两种分析方法所得结果与GC-MS特征组分分析的结果进行比对发现,Fisher判别能够对7种助燃剂燃烧残留物的样本实现更有效的分类,对未知样本的判别更为有效。该研究结果为火场助燃剂鉴定提供了新的数据分析手段。     

Analysis of t-Butyldimethylsilyl Derivatives of Chlorophenols in the Atmosphere of Urban and Rural Areas in East of France

A method using GC-MS and derivatization with N-(t-butyldimethylsilyl)-N-ethyltrifluoroacetamide (MTBSTFA) was developed for the analysis of 19 chlorophenols compounds in atmospheric samples (gas and particles). Air sampling was carried out using a Hi-Vol sampler with glass fibre filter and XAD-2 resin at a flow rate of 60 m³ h⁻¹. The particle and gas phases were collected separately over a period of 4 h. Samples were Soxhlet extracted, evaporated to dryness under nitrogen and refilled with acetonitrile. 100 mL of these extracts were derivatized with 100 μL of MTBSTFA at 80 °C for 1 h under strong stirring. Sylylated chlorophenols were injected into a GC-MS in splitless mode and quantified as their TBDMS derivatives in the SIM mode. Mass spectral analysis of the derivatives of the 19 compounds studied indicates that the spectra are highly specific showing an ion at [M - 57]⁺ which is useful for structure confirmation or analysis at low levels using selected ion monitoring. Quantification limits varied between 5 μg L⁻¹ and 10 μg L⁻¹ which correspond to 20 pg m⁻³ and 40 pg m⁻³ for 250 m³ of air sampled. This method was successfully applied to atmospheric samples collected simultaneously in winter 2004 in an urban (Strasbourg) and rural (Erstein) areas in east of France.     

Validation of a confirmatory method for the determination of melamine in egg by gas chromatography-mass spectrometry and ultra-performance liquid chromatography-tandem mass spectrometry

A sensitive and reliable method was developed and validated for detection and confirmation of melamine in egg based on gas chromatography-mass spectrometry (GC-MS) and ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Trichloroacetic acid solution was used for sample extraction and precipitation of proteins. The aqueous extracts were subjected to solid-phase extraction by mixed-mode reversed-phase/strong cation-exchange cartridges. Using ultra-performance liquid chromatography and electrospray ionization in the positive ion mode, melamine was determined by LC-MS/MS, which was completed in 5 min for each injection. For the GC-MS analysis, extracted melamine was derivatized with N,O-bis(trimethylsilyl)trifluoracetamide prior to selected ion monitoring detection in electron impact mode. The average recovery of melamine from fortified samples ranged from 85.2% to 103.2%, with coefficients of variation lower than 12%. The limit of detection obtained by GC-MS and UPLC-MS/MS was 10 and 5 μg kg⁻¹, respectively. This validated method was successfully applied to the determination of melamine in real samples from market.     

Supervised pattern recognition in food analysis

Data analysis has become a fundamental task in analytical chemistry due to the great quantity of analytical information provided by modern analytical instruments. Supervised pattern recognition aims to establish a classification model based on experimental data in order to assign unknown samples to a previously defined sample class based on its pattern of measured features. The basis of the supervised pattern recognition techniques mostly used in food analysis are reviewed, making special emphasis on the practical requirements of the measured data and discussing common misconceptions and errors that might arise. Applications of supervised pattern recognition in the field of food chemistry appearing in bibliography in the last two years are also reviewed.     

Comparison of Volatile Organic Compounds of Sideritis romana L. and Sideritis montana L. from Croatia

A study on the headspace volatile organic compounds (VOCs) profile of native populations of Sideritis romana L. and Sidertis montana L., Lamiaceae, from Croatia is reported herein, to elucidate the phytochemical composition of taxa from this plant genus, well-known for traditional use in countries of the Mediterranean and the Balkan region. Headspace solid-phase microextraction (HS-SPME), using divinylbenzene/carboxene/polydimethylsiloxane (DVB/CAR/PDMS) or polydimethylsiloxane/divinylbenzene (PDMS/DVB) fiber, coupled with gas chromatography-mass spectrometry (GC-MS) was applied to analyze the dried aerial parts of six native populations in total. Furthermore, principal component analysis (PCA) was conducted on the volatile constituents with an average relative percentage ≥1.0% in at least one of the samples. Clear separation between the two species was obtained using both fiber types. The VOCs profile for all investigated populations was characterized by sesquiterpene hydrocarbons, followed by monoterpene hydrocarbons, except for one population of S. romana, in which monoterpene hydrocarbons predominated. To our knowledge, this is the first report on the VOCs composition of natural populations of S. romana and S. montana from Croatia as well as the first reported HS-SPME/GC-MS analysis of S. romana and S. montana worldwide.     

Self organising maps for visualising and modelling

The paper describes the motivation of SOMs (Self Organising Maps) and how they are generally more accessible due to the wider available modern, more powerful, cost-effective computers. Their advantages compared to Principal Components Analysis and Partial Least Squares are discussed. These allow application to non-linear data, are not so dependent on least squares solutions, normality of errors and less influenced by outliers. In addition there are a wide variety of intuitive methods for visualisation that allow full use of the map space. Modern problems in analytical chemistry include applications to cultural heritage studies, environmental, metabolomic and biological problems result in complex datasets. Methods for visualising maps are described including best matching units, hit histograms, unified distance matrices and component planes. Supervised SOMs for classification including multifactor data and variable selection are discussed as is their use in Quality Control. The paper is illustrated using four case studies, namely the Near Infrared of food, the thermal analysis of polymers, metabolomic analysis of saliva using NMR, and on-line HPLC for pharmaceutical process monitoring.     

Four levels of pattern recognition

Problems of pattern recognition in chemistry and other subjects can be divided conveniently into four different types depending on the level of scope of the problem.(1) Classification into one of a number of defined classes. As an example blood samples taken from persons known to be either controls or welders are considered. The problem is whether trace element concentrations in these samples contain information on whether or not a person is a welder.(2) Level 1 plus the possibility that an object is an outlier, i.e. does not belong to any of the defined classes. As an example, the üse of ¹³C-n.m.r. data to decide whether 2-substituted norbornanes have the exo or endo structure is discussed. (2A) Level 2, asymmetric. This situation occurs when one class does not have a systematic structure, but another class is homogeneous and can be described by a level 2 model. This occurs in the classification of materials or compounds as good or bad, active or inactive, and in binary classifications. As an example the use of trace element data to classify steel samples as having good or poor properties of strength is discussed.(3) Level 2 plus the ability to relate the variables measured to external properties of continuous character. As an example, the classification of a series of chemical compounds as β -receptor blockers, β -receptor stimulants, or neither, on the basis of their structural variables is discussed. In addition, relations between these structural variables and the measured biological activity are sought within each of the two classes.(4) Level 3 with the difference that several external property variables in the objects are measured. It may be desirable to use variables of the objects both for classification and for relations to several property variables: such examples are numerous in analytical chemistry.     

Comparing roadsoils pollution patterns extracted by MOLMAP and classical three-way decomposition methods

A recent approach based on self-organizing maps (SOMs) to extract patterns from three-way data, named MOLMAP, was applied in a four-seasons study on soil pollution and its results compared with three different conventional approaches: Parallel factor analysis (PARAFAC), matrix augmented principal components analysis (MA-PCA) and Procrustes rotation. Each sampling season comprised 92 roadsoil samples and 12 analytical variables (Cd, Co, Cu, Cr, Fe, Mn, Ni, Pb, Zn, loss on ignition, pH and humidity). It was found that all techniques yielded highly similar results as the samples became organized in two major groups, each with a differentiated pollution pattern. This confirmed MOLMAP as a reliable option to handle environmental three-way datasets and to extract accurate pollution patterns.     

Application of Alternative K-Nearest Neighbor method to Classification of Alloy Samples

Pattern Recognition is one of main methods that can resolve structure of chemistry and classification of compounds. In chemistry, K-Nearest Neighbor method (KNN) is mostly used for recognition of compounds that are not classified by a linear method. ALKNN (Alternative KNN) is an improvement of KNN. The method is based on the classification rules which are produced by using standards and knowledge. The unknown samples are,then, classified with the aid of the rules.     

One step derivatization with British Anti-Lewsite in combination with gas chromatography coupled to triple-quadrupole tandem mass spectrometry for the fast and selective analysis of inorganic arsenic in rice

We developed a new fast and selective analytical method for the determination of inorganic arsenic (iAs) in rice by a gas chromatography – tandem mass spectrometry (GC-MS/MS) in combination with one step derivatization of inorganic arsenic (iAs) with British Anti-Lewsite (BAL). Two step derivatization of iAs with BAL has been previously performed for the GC-MS analysis. In this paper, the quantitative one step derivatization condition was successfully established. The GC-MS/MS was carried out with a short nonpolar capillary column (0.25 mm × 10 m) under the conditions of fast oven temperature ramp rate (4 °C/s) and high linear velocity (108.8 cm/s) of the carrier gas. The established GC-MS/MS method showed an excellent linearity (r² > 0.999) in a tested range (0.2–100.0 μg L⁻¹), ultra-low limit of detection (LOD, 0.08 pg), and high precision and accuracy. The GC-MS/MS technique showed far greater selectivity (22.5 fold higher signal to noise ratio in rice sample) on iAs than GC-MS method. The gas chromatographic running time was only 2.5 min with the iAs retention time of 1.98 min. The established method was successfully applied to quantify the iAs contents in polished rice. The mean iAs content in the Korean polished rice (n = 27) was 66.1 μg kg⁻¹ with the range of 37.5–125.0 μg kg⁻¹. This represents the first report on the GC-tandem mass spectrometry in combination with the one step derivatization with BAL for the iAs speciation in rice. This GC-MS/MS method would be a simple, useful and reliable measure for the iAs analysis in rice in the laboratories in which the expensive and element specific HPLC-ICP-MS is not available.