In many applications of neural networks, e.g. time series prediction or pattern analysis, training data are generated automatically out of large data sets. The problem is to determine the varying significance of the resulting training vectors concerning the given task in order to make appropriate decisions for the training phase. In this paper we propose a self-organized significance analysis based on a rareness assessment for each vector in the generated training data set. The resulting significance measure can be used to achieve considerably improved classification results for a wide variety of applications by systematically controlling training parameters like learning rate or frequency of presentation for each single vector.

High tech (HT) trace elements such as germanium, gallium and indium gain rising importance in the development of innovative technologies. The database "HTMET" forms the first nationwide metal-ore database for Germany, created to visualise HT metal characteristics of base metal ores from important mining districts. Mineralogical and geochemical investigations on 478 samples and ore concentrates from 109 Pb-Zn-Cu occurrences were carried out using analytical methods with high spatial resolution and bulk sample methods. The database provides aggregated data based on 17,000 geochemical data sets, compiled information on regional infrastructure and environmental risks as well as data on innovative raw material-efficient processing techniques. Evaluation of combined data provides interactive maps revealing new potentials for specific HT metals in Germany. Differences in regional distribution of these trace elements and dependency of their concentration levels in the ore on the genetic deposit type became apparent. Sphalerite from the sediment-hosted massive sulphide (SHMS) deposit "Rammelsberg" and skarn deposits in the Erzgebirge contain elevated indium contents (median 14-119 ppm), whereas the SHMS deposit "Meggen" is poor in HT metals. Germanium forms the predominant HT trace element in colloform sphalerite of Mississippi-Valley-Type (MVT) deposits (median 29-147 ppm); in contrast, crystalline sphalerite is low in germanium in this deposit type. Sphalerite in all hydrothermal vein deposits shares a distinct enrichment in gallium (median 6-81 ppm); however, germanium and indium concentrations vary significantly depending on the metal source and fluid conditions. The Ruhrgebiet and the Schwarzwald ore veins show an enrichment in germanium (median 55-73 ppm), whilst vein sphalerite from the Erzgebirge is specialised in indium (median 33 ppm). The data demonstrate that the HT trace element inventory of the studied base metal sulphides is not only a function of the genetic ore deposit type, but is also triggered by locally variable geology such as source rock and fluid composition and organic content of the rock. Gallium seems to derive from adjacent lithologies, whereas indium and germanium may have more distant sources. Kurzfassung: Durch die fortschreitende Verbreitung und Weiterentwicklung von modernen Technologien erlangen die sogenannten hochtechnologierelevanten (HT) Spurenelemente Germanium, Gallium und Indium steigende Bedeutung. Die "HTMET"-Datenbank ist das erste deutschlandweite Metallerz-Kataster, welches die HT-Metall-"Spezialisierung" wichtiger Buntmetall-Bergbaudistrikte reflektiert. An 478 Erzproben und-konzentraten aus 109 Pb-Zn-Cu-Vorkommen wurden mittels hochortsaufgelösten Punktanalysen und nasschemischen Gesamterzmessungen mineralogische und geochemische Untersuchungen durchgeführt. Die Datenbank beinhaltet aggregierte Daten von 17.000 geochemischen Datensätzen, Informationen zur regionaler Infrastruktur und möglichen Umweltrisiken, sowie Ergebnisse aus innovativen, rohstoffeffizienten Aufbereitungs...

No abstract

No abstract