Finding the best visualization of an ontology

An ontology is a classification model for a given domain. In information retrieval, ontologies are used to perform broad searches. An ontology can be visualized as nodes and edges. Each node represents an element and each edge a relation between a parent and a child element. Working with an ontology becomes easier with a visual representation. An idea is to use the expressive power of 3D representation to provide visualization for the user. In this paper, we propose a new method for positioning the elements of the visualized concept lattice in the 3D world based on operations research (OR) methods. One method uses a discrete location model to create an initial solution and we propose heuristic methods to further improve the visual result. We evaluate the visual results according to our success criteria and the feedback from users. Running times of the heuristic indicate that an improved version should be feasible for on-line processing and what-if analysis of ontologies.     

Isotopic study of the mechanism of ozone formation

The mechanism of ozone formation has been studied using ¹⁶O and ¹⁸O₂. High-resolution microwave spectroscopy was used to measure the amounts of the isotopomeric ozone species formed. The study is hampered by the very rapid exchange process between the reactants, that tends to scramble the isotopes and hence give a 2:1 statistical ratio between the two possible isotopomers. We have found a strategy to come around this difficulty and conclude that the mechanism is a simple end-on-addition.     

Improved customer choice predictions using ensemble methods

In this paper various ensemble learning methods from machine learning and statistics are considered and applied to the customer choice modeling problem. The application of ensemble learning usually improves the prediction quality of flexible models like decision trees and thus leads to improved predictions. We give experimental results for two real-life marketing datasets using decision trees, ensemble versions of decision trees and the logistic regression model, which is a standard approach for this problem. The ensemble models are found to improve upon individual decision trees and outperform logistic regression.     

Working Methods Paper: Critical considerations with respect to the identification of tin species in the environment

Biogeochemical pathways of tin species in the environment are still controversial, e.g. with regard to methylation and transmethylation phenomena, owing to the fact that the identification of methylated tin-compounds is often difficult. The previous tentative identification of a mixed methylbutyltin compound in sediment and biological samples by GC/AAS after hydride generation gave an illustration of this problem. This compound was previously identified in sediments by other authors and also suspected to occur in a contaminated sediment sample from the Boyardville Marina, France. The retention time obtained by GC/AAS corresponded to the actual retention time of a mixed methylbutyltin calibrant. However, additional checks demonstrated that the compound detected was actually monophenyltin. This evidence was produced by a thorough analysis of a selected sediment sample by alternative techniques such as GC/AAS and GC/AES after pentylation, GC/FPD and GC/MS. The results presented highlight the need for a full identification of compounds to avoid mis-interpretation.