Globally Optimal Parsimoniously Lifting a Fuzzy Query Set Over a Taxonomy Tree
The paper defines an annotated suffix tree (AST) - a data structure used to calculate and store the frequencies of all the fragments of the given string or a collection of strings. The AST is associated with a string to text scoring, which takes all fuzzy matches into account. We show how the AST and the AST scoring can be used for Natural Language Processing tasks. Copyright © by the paper's authors. Copying only for private and academic purposes.
Abstract Post-conflict affiliation between former opponents and bystanders occurs in several species of non-human primates. It is classified in four categories of which affiliation received by the former victim, ‘consolation’, has received most attention. The hypotheses of cognitive constraint and social constraint are inadequate to explain its occurrence. The cognitive constraint hypothesis is contradicted by recent evidence of ‘consolation’ in monkeys and the social constraint hypothesis lacks information why ‘consolation’ actually happens. Here, we combine a computational model and an empirical study to investigate the minimum cognitive requirements for post-conflict affiliation. In the individual-based model, individuals are steered by cognitively simple behavioural rules. Individuals group and when nearby each other they fight if they are likely to win, otherwise, they may groom, especially when anxious. We parameterize the model after empirical data of a tolerant species, the Tonkean macaque (Macaca tonkeana). We find evidence for the four categories of post-conflict affiliation in the model and in the empirical data. We explain how in the model these patterns emerge from the combination of a weak hierarchy, social facilitation, risk-sensitive aggression, interactions with partners close-by and grooming as tension-reduction mechanism. We indicate how this may function as a new explanation for empirical data.
We describe a novel method for the analysis of research activities of an organization by mapping that to a taxonomy tree of the field. The method constructs fuzzy membership profiles of the organizationmembers or teams in terms of the taxonomy’s leaves (research topics), and then it generalizes them in two steps. These steps are: (i) fuzzy clustering research topics according to their thematic similarities in the department, ignoring the topology of the taxonomy, and (ii) optimally lifting clusters mapped to the taxonomy tree to higher ranked categories by ignoring “small” discrepancies. We illustrate the method by applying it to data collected by using an in-house e-survey tool from a university department and from a university research center. The method can be considered for knowledge generalization over any taxonomy tree.