A neuro-fuzzy system for X-ray spectra interpretation

A decision scheme for the interpretation of spectra from wavelength dispersive X-ray fluorescence spectrometry is described that encompasses elements from three areas of artificial intelligence: fuzzy logic, rule based expert systems and neural net technology.After transforming the recorded spectra to line spectra by appropriate background correction a reasoning scheme is applied that takes into account not only the observed spectra, but also the recording conditions and prior spectroscopic information regarding the relative emission probabilities and the usefulness of the different lines for the purpose of element identification. The latter is done on the basis of a previously described scheme to compute conditional a posteriori Bayes probabilities for a mean matrix. These different pieces of information are then assembled into a battery of fuzzy rules. The importance of the rules as well as the importance of the X-ray lines is determined in a training process, similar to the one in a feedforward back-propagation network.To further stabilize the results this network is pruned in a second training cycle. This, however, had little effect on the quality of interpretation.The advantages of this approach to the interpretation of X-ray spectra over older ones are numerous: the system adapts itself to better interpret spectra that are of greater importance to a laboratory as these are better represented in the training set; the fuzzy logic is capable of working with incomplete and uncertain knowledge, and the neural network results based on these fuzzy rules is readily interpretable by the X-ray spectroscopist as every rule can be expressed also in natural language as in any classical rule based system.On leave from Silesian University, Katowice, Poland     

A fundamental parameters approach including scattered radiation for mono-energetically excited samples in energy-dispersive x-ray fluorescence spectrometry

A fundamental parameters approach that incorporates data on scattered radiation is shown to be a useful tool for the determination of trace elements in light matrices. For thick, transparent samples, a computer program is developed that uses differetial intensity equations adjusted for decreasing geometric efficiency. Various organic and biological reference materials are analysed to assess the reliability of the procedure. A nested design for analysis of variance shows that major errors can originate from the deconvolution process and sample inhomogeneities in many cases. For loose powders, great care must be exercised to avoid errors caused by scatter from the sample cups. The program is also useful for alloys.     

Quality assurance in environmental research: emissions and fate of platinum

Summary Quality assurance in environmental research is related to reliable analytical methods. According to this fact great efforts were taken in the German Cooperative Project sponsored by the Ministry of Research and Technology Emissions of Precious Metals to ensure the quality of all analytical data from various research groups. As there are only few certified reference materials (CRMs) for platinum available and the application of reference materials is an important measure in quality assurance and quality control actions in every laboratory, several materials have been prepared with low and elevated platinum contents as reference samples for this project. These reference samples were a catalyst sample with elevated platinum content, a dust reference material with low platinum content and a catalyst reference material with low platinum content. With these reference samples and the Canadian geological reference material SU-1a interlaboratory studies were conducted and the results of these are presented in this paper.Dedicated to Professor Dr. Wilhelm Fresenius on the occasion of his 80th birthday     

Single- and multi-channel detection for generalized quantitative analysis in cases of unresolved chromatographic peaks

Computerized quantification of components under overlapping chromatographic peaks is done by calibration of chromatograms against component mixtures. For conventional (single-channel) detectors, the limitations of earlier methods based on ordinary multiple regression, can be circumvented by data reduction with the aid of principal component analysis with the partial least-squares approach. Simulation studies show that the method can be applied even when there is severe peak overlap, unstable baseline, noisy chromatograms or non-linear detector response. Advantages in the quantification of fused peaks by means of multichannel detectors are outlined. Present limitations on the quantitative evaluation of several overlapping component peaks from a single spectro-chromatogram by means of the partial least-squares method combined with multiple regression on the pure component spectra, are discussed with respect to practical high-performance liquid chromatography.     

Exact sequence analysis for three-dimensional hydrophobic-polar lattice proteins

We have exactly enumerated all sequences and conformations of hydrophobic-polar (HP) proteins with chains of up to 19 monomers on the simple cubic lattice. For two variants of the HP model, where only two types of monomers are distinguished, we determined and statistically analyzed designing sequences, i.e., sequences that have a nondegenerate ground state. Furthermore we were interested in characteristic thermodynamic properties of HP proteins with designing sequences. In order to be able to perform these exact studies, we applied an efficient enumeration method based on contact sets.     

Polymer-attractive spherical cage system

We analyze the structural behavior of a single polymer chain inside an attractive sphere. Our model is composed of a coarse-grained polymer governed by Lennard-Jones interactions of the monomers and an attractive sphere potential which follows by integrating the monomer-monomer interaction over the (inner) surface of the sphere. By means of extensive multicanonical Monte Carlo simulations it is shown that the system exhibits a rich phase diagram in the adsorption strength-temperature (ε ? T) plane ranging from highly ordered and compact to extended and random coil structures and from desorbed to partially or even completely adsorbed conformations. These findings are identified with different energetic and structural observables. The resulting phase diagram in the ε ? T plane is compared with that for a polymer adsorbing to a plane, attractive substrate obtained previously by Möddel, Bachmann, and one of the authors.     

Multicanonical chain-growth algorithm

We present a temperature-independent Monte Carlo method for the determination of the density of states of lattice proteins that combines the fast ground-state search strategy of the new pruned-enriched Rosenbluth chain-growth method and multicanonical reweighting for sampling the complete energy space. Since the density of states contains all energetic information of a statistical system, we can directly calculate the mean energy, specific heat, Helmholtz free energy, and entropy for all temperatures. We apply this method to lattice proteins consisting of hydrophobic and polar monomers, and for the examples of sequences considered, we identify the transitions between native, globule, and random coil states. Since no special properties of heteropolymers are involved in this algorithm, the method applies to polymer models as well.     

Traceability without uncertainty: current situation in the pharmaceutical industry

The current situation in the pharmaceutical industry is discussed, when the traceability of measurement (analytical) results to certified values of pharmacopoeial reference standards is required, without evaluating their uncertainties. It is shown that the evaluation of measurement uncertainty is necessary for understanding the level of confidence of the analytical results and their comparability, particularly during preparation and characterisation of the reference standards.Papers published in this section do not necessarily reflect the opinion of the Editors, the Editorial Board and the Publisher. Apart from exceptional circumstances, they are not submitted to the usual referee procedure and go essentially unaltered.     

Modelling and Estimation of Measurement Errors

Accreditation and Quality Assurance -