Classification of ancient Roman glazed ceramics using the neural network of Self-Organizing Maps

Artificial neural networks with unsupervised learning strategy known as Self-Organizing Maps were applied to classify ancient Roman glazed ceramics. Their clay ceramic bodies were analyzed by Inductively Coupled Plasma-Atomic Emission Spectroscopy and the chemical composition obtained was processed by this neural algorithm. The results obtained provide two types of information: firstly, classification of ceramic samples with identification of several groups and secondly, differentiation between the elemental chemical information. It was found that there are certain chemical elements which can be considered as principal and which can serve to differentiate between ceramics, whereas other elements give redundant information and do not contribute to sample differentiation. Seven chemical elements were considered principal and provide the necessary information. Two types of element were identified: 1- a group formed by common elements, such as: Ca, Fe, Mg, Mn and 2- another formed by optional elements: K or Na and Ba or Sr and Al or Ti.     

用人工神经网络方法对多环芳烃的致癌性进行分类

将自适应映射（ＳＯＭ）用于多环芳烃致癌性的分级。采用的输入参数为分子比表面积、代谢活性区及亲电活性区的中心碳原子离域能,分子中脱毒区总数。优化的网络参数包括网络数及网格形状,学习次数和学习率、邻居半径等。在最佳网络参数下,多环芳烃致癌性分类准确度大于９７％。     

EDXRF imaging of Pb in glazed ceramics using a micropattern gas detector

A new system for energy-resolved X-ray fluorescence imaging using a microhole and strip plate (MHSP), a new type of micropattern gas detector (MPGD), is proposed. It works as a single photon counting detector with position and energy detection capability. The interaction of X-rays with the gas medium produces electrons via the photoelectric effect, and the number of electrons is proportional to the absorbed X-ray energy. These electrons are further multiplied in the MHSP. Position detection is achieved using the charge division method. The detector has an active area of 28?×?28 mm² and shows good position resolution, about σ?=?125 μm, an intrinsic energy resolution of about 14% FWHM for 5.9 keV X-rays, and a counting rate capability of up to 0.5 MHz. The system has shown good properties for energy-dispersive X-ray fluorescence (EDXRF) applications, since it allows efficient energy and position detection of fluorescence X-rays from multielemental samples. In this work, the system was used to study lead depth distributions in eighteenth-century Portuguese faiences from the Santa Clara-a-Velha monastery. The fluorescence images were obtained by irradiating the samples, with a pinhole placed between the sample and the detector to focus the radiation into the detector. The results are presented here, including the elemental map distributions for different samples.     

Provenancing of ancient Roman millstones

Twenty samples of millstones made from volcanic lavarecovered from two Roman villa sites in southern Italy, covering the period IInd/Ist century BC, have been analyzed by instrumental neutron activation using the SLOWPOKE reactor at the University of Toronto. Also included in the studies were three samples of lava from an extinct volcano located in the same region. Findings are compared with published data on trace element compositions from a number of Mediterranean volcanic sources. Results show the provenance of the millstones to be from two different Italian sources.     

Thermogravimetric investigation of ancient ceramics

Sherds from restored ancient pots taken from archaeological sites of Siberian region (Late Bronze and Early Iron Age, IX–VIII to VII–VI BC) were investigated by thermogravimetry in order to define the effects of sampling. Three types of the sampling were (1) scanning through the inner surface of a pot, (2) outer surface, core, and inner surface of thick-walled sherds, and (3) random fragments of a restored pot. The results of the measurements were shown to depend on two factors, clay paste composition and firing conditions. Redistribution between mass loss at dehydration and dehydroxylation was detected for the ancient ceramics after ‘mild’ firing. The results of the measurements are explained in terms of a temperature profile throughout the wall of a pot during the thermal treatment under firing and cooking meal.     

The mass-loss diagram for the ancient ceramics

The procedure for the visualization of thermogravimetric results for ancient ceramics is described and justified. Two main parameters for the analysis are dehydration and dehydroxylation, which are derived from the experiment as the values of mass loss at temperature interval from room temperature to 350 °C and from 350 to 600 °C. Three examples show how to use the mass-loss diagram for (1) the analysis of the material of a single pot; (2) comparison between different pots from the same archeological site, and (3) search for the source of clay for the manufacturing ancient ceramics.     

Thermal analysis of Romanian ancient ceramics

The present work is focused on thermoanalytical investigations as thermogravimetric analysis (TG) and derivative thermal analysis (DTG), applied for the characterization of some samples collected from archaeological sites (Brasov and Trofeum Traiani) located in different regions of Romania. New informations derived about ceramic technologies concerning raw materials and binding materials (mineralogical components) have been obtained. All these experimental results have been correlated with related techniques as X-ray diffraction (XRD), energy-dispersive X-ray fluorescence (EDXRF) and inductively coupled plasma—atomic emission spectrometry (ICP-AES). By progressive heating in static air atmosphere and in the temperature range of 20–800 °C, all investigated materials exhibit three main successive processes, associated with the dehydration and thermo-oxidative degradations. The rate of the first thermooxidative process, temperatures corresponding to the maximum rate of the second thermooxidative process and shrinkage temperature were associated with the damage of the investigated materials due to environmental impact. Heating also affects the contact between the fine-sized clay matrix and mineral clast fragments, appearing in reaction rims, sometimes showing newly formed phases. The temperature at which ancient ceramics and pottery were fired varies over a wide range (600–800 °C) depending on the type of clay used, although firing temperatures not above 30–400 °C have also been suggested. Clay minerals, as the main material for production of ceramics and pottery, show some characteristic reactions (dehydroxylation, decomposition, transformation) in the course of firing (heating effects) and several thermoanalytical criteria can be used for reconstruction of former production conditions.     

Classification of ecstasy tablets using trace metal analysis with the application of chemometric procedures and artificial neural network algorithms

This work is concerned with an investigation into the practicalities of using ICP-MS data obtained from the analysis of ecstasy tablets to provide linkage information from seizure to seizure. The generated data was analysed using different statistical techniques, namely principal component analysis, Hierarchical clustering and artificial neural networks. The relative merits of these different techniques are discussed.     

Undergraduate Projects in the Application of Artificial Intelligence to Chemistry. II Self-Organizing Maps

It is often necessary in science to identify samples that have features in common. For example, one might wish to find those NMR spectra in a large database that have similar patterns of resonances or identify samples amongst a large number of specimens of river water that analysis shows have similar biochemical oxygen demand, heavy metals concentration, organochlorine content, and so on.The determination of relationships among samples is a task to which Artificial Intelligence is increasingly being applied. In this paper, we investigate the Self-Organizing Map (SOM), whose role is to perform just this kind of task; in other words, to cluster data samples so as to reveal the relationships that exist among them. The self-organizing map is a method, which, unusually, combines a mathematical foundation with an intuitive interpretation.We will describe how a simple SOM operates, what kinds of data may be analyzed using one, and how a computer program to run a SOM can be written by anyone-whether student or teacher-with modest programming skills. Portions of sample source code are included in this paper, and program listings for the examples that are discussed are available in the supporting materials. The supporting files can also be used to see the maps in operation.     

An attractive way to recover information from ancient ceramics using EPR spectroscopy

Electron Paramagnetic Resonance (EPR) spectroscopy is useful to archeologists in recognizing and dating sites. This non-destructive technique of analysis gives information on material composition, on paramagnetic impurities, on firing techniques and on surface oxidation. In this paper it is shown how to recover information about firing techniques from a fragment of ceramic pot, dating from the 9th century B.C. The internal brown-black layer showed a strong g anisotropy whilst the external reddish layer showed hardly any. This indicates a non-homogeneous rapid firing of the pot, due probably to primitive firing techniques.     

Classification and identification of mosquito species using artificial neural networks

An artificial neural network method is presented for classification and identification of Anopheles mosquito species based on the internal transcribed spacer2 (ITS2) data of ribosomal DNA string. The method is implemented in two different multi-layered feed-forward neural network model forms, namely, multi-input single-output neural network (MISONN) and multi-input multi-output neural network (MIMONN). A number of data sequences in varying sizes of different Anopheline malarial vectors and their corresponding species coding are employed to develop the neural network models. The classification efficiency of the network models for untrained data sequences is evaluated in terms of quantitative performance criteria. The results demonstrate the efficiency of the neural network models to extract the genetic information in ITS2 sequences and to adapt to new data. The method of MISONN is found to exhibit superior performance over MIMONN in distinguishing and identification of the mosquito vectors.     

基于自组织特征映射神经网络的金银花分类研究

自组织特征映射神经网络(SOM)以无监督方式进行网络训练,具有自组织功能。网络通过自身训练,自动对输入模式进行分类。中药药用价值与其所含微量元素有直接的关系,药材分类是中药质量控制的重要方法。将金银花中微量元素含量作为网络输入,利用自组织特征映射神经网络对不同产地金银花进行分类。结果表明分类效果较好,符合生产实际。     

Monitoring of yeast fermentation by ion mobility spectrometry measurement and data visualisation with Self-Organizing Maps

Ion mobility spectrometry (IMS) measurement combined with unsupervised neurocomputing is considered as a new potential method for on-line monitoring of fermentation and other processes producing volatile compounds that involve micro-organisms. This was demonstrated in a model system in which a strain of brewer’s yeast (Saccharomyces cerevisiae) was cultivated in a bench-top fermenter. Five phases of yeast growth could be detected from measurements of the exhaust gases from the fermenter, as indicated by the changes in ion mobility spectra analysed by computational methods.The data were first processed using the Self-Organizing Map (SOM) algorithm, the results showing that the phases of fermentation can be detected and identified. The cultivations were also shown by Sammon’s mapping to be comparable to a certain level of accuracy. Contaminated cultivation could be detected by its distinctive ion mobility spectrometry profile.     

Analysis of ultrasonic velocity in refrigerants using artificial neural network

Several researchers have reported numerous measurements on ultrasonic velocity as a function of temperature and pressure using various experimental techniques. A large amount of experimental data is required in order to obtain accurate results for the chemical substances used. The present article explores the evaluation of ultrasonic velocity as a function of molecular weight, temperature and pressure using an artificial neural network (ANN) in six refrigerants. The network so developed predicts the ultrasonic velocity successfully. Statistical analysis of the results was performed using standard deviation (%) and relative average deviation. The correlation coefficient in our analysis was found to be 0.9999. The trained weights, obtained from ANN, are further employed to form equations to predict ultrasonic velocity at other temperatures and pressures.     

Self-Organizing Maps and Support Vector Regression as aids to coupled chromatography: illustrated by predicting spoilage in apples using volatile organic compounds

The paper describes the application of SOMs (Self-Organizing Maps) and SVR (Support Vector Regression) to pattern recognition in GC-MS (gas chromatography-mass spectrometry). The data are applied to two groups of apples, one which is a control and one which has been inoculated with Penicillium expansum and which becomes spoiled over the 10-day period of the experiment. GC-MS of SPME (solid phase microextraction) samples of volatiles from these apples were recorded, on replicate samples, over time, to give 58 samples used for pattern recognition and a peak table obtained. A new approach for finding the optimum SVR parameters called differential evolution is described. SOMs are presented in the form of two-dimensional maps. This paper shows the potential of using machine learning methods for pattern recognition in analytical chemistry, particularly as applied to food chemistry and biology where trends are likely to be non-linear.