Neuro-fuzzy approach to processing inputs with missing values in pattern recognition problems

An approach to dealing with missing data, both during the design and normal operation of a neuro-fuzzy classifier is presented in this paper. Missing values are processed within a general fuzzy min–max neural network architecture utilising hyperbox fuzzy sets as input data cluster prototypes. An emphasis is put on ways of quantifying the uncertainty which missing data might have caused. This takes a form of classification procedure whose primary objective is the reduction of a number of viable alternatives rather than attempting to produce one winning class without supporting evidence. If required, the ways of selecting the most probable class among the viable alternatives found during the primary classification step, which are based on utilising the data frequency information, are also proposed. The reliability of the classification and the completeness of information is communicated by producing upper and lower classification membership values similar in essence to plausibility and belief measures to be found in the theory of evidence or possibility and necessity values to be found in the fuzzy sets theory. Similarities and differences between the proposed method and various fuzzy, neuro-fuzzy and probabilistic algorithms are also discussed. A number of simulation results for well-known data sets are provided in order to illustrate the properties and performance of the proposed approach.     

Light-induced starch degradation in non-dormant turions of Spirodela polyrhiza

Continuous red light controls starch degradation in turions of Spirodela polyrhiza[Dölger, K., U. K. Tirlapur and K.‐J. Appenroth [1997]Photochem. Photobiol. 66, 126–127 (1997)]. This light could be replaced by repeated red light pulses with the reciprocity law fulfilled over a large range of fluence rates. The effect of red light pulses repeated every 24 or 12 h for 6 days was reversible by subsequent far‐red light pulses. In contrast, hourly applied red pulses were irreversible by far‐red light. This discrepancy was explained by showing the starch degradation activity of far‐red pulses themselves. The investigated process was categorized as a phytochrome low fluence response with an unusual property: requirement of light treatment for several days. A partial fulfillment of this requirement was obtained with a red pulse followed by a dark period and a 24 h continuous irradiation. These results suggest the existence of two separate steps in the process of starch degradation in turions: formation of a sprout (=sink) during the pulse‐induced germination, and starch degradation in the storage tissue (=source) induced by the second light treatment.     

Some Isoperimetric Inequalities in Electrochemistry and Hele Shaw Flows

Isoperimetric inequalities are applied to a moving-boundaryproblem for doubly-connected domains. This problem occurs forexample in electrochemistry, in which case the domains in questionare the electrolyte of an electrolytic cell. The two electrodessurrounding the electrolyte are assumed to grow or dissolve,at different rates in general, by electrochemical reaction.We obtain optimal estimates showing, for example, that the leastchange in volume of each electrode always occurs in sphericalsymmetry.     

Dynamic anisotropy and heterogeneity of polystyrene thin films as studied by inelastic neutron scattering

We studied the dynamic anisotropy and heterogeniety of polystyrene thin films in glassy state with inelastic neutorn scattering. Adjusting the scattering vector to the directions parallel and perpendicular to the film surface, we observed the elastic scattering intensities as a function of temperature. It was found for the 200 A film that the elastic intesity decreased with increasing temperature more rapidly in the perpendicular direction than in the pararell direction, showing the higher mobility in the perpendicular direction. However, such dynamical anisotropy was not observed in the 1000 A film. The decrease in the mobility was observed with the film thickness in both the directions. These results were explained in terms of an interface hard layer. We also evelauated the dynamical heterogeniety from the non-Gaussian parameter A0, which increased with decreasing the film thickness, showing the increase in the dynamical heterogeneity. Assuming a simple bi-layer model consisting of the interface hard layer and the bulk-like layer, we analyzed the thickness dependence of the non-Gaussian parameter A0 and the mean square displacement (u2) to find that the hard layer has a thickness of approximately 130 A and a mean square displacement of approximately 0.018 A2 at 230 K.