Kernels are problem-specific functions that act as an interface between the learning system and the data. While it is well-known when the combination of two kernels is again a valid kernel, it is an open question if the resulting kernel will perform well. In particular, in which situations can a

in performance between various state-of-the-art texture descriptors proves rather small in this task. On the other hand, we clearly demonstrate that very significant gains can be achieved via different SVM-based classification techniques. Selecting appropriate kernel parameters proves crucial. This motivates a

a kernel can be defined. This approach is shown to give state of the art categorisation performance.

The accurate recognition of modal information is vital for the correct interpretation of statements. In this paper, we report on the collection a list of words and phrases that express modal information in biomedical texts, and propose a categorisation scheme according to the type of information

transformations (topological or morphing transformations). Previous experiments with line drawings showed that categorisation performance depends systematically on the amount of shape transformation, both in speeded categorisation tasks and rating tasks (Graf, 2002). We investigated whether this systematic

performance indicators to measure the productivity of modular services using an IT tool

In this article we examine three different ap-proaches to categorising documents from mul-tilingual corpora using machine learning algo-rithms. These approaches satisfy two main conditions: there may be an unlimited num-ber of different languages in the corpus and it is unnecessary to previously

structure with no feedback during the task. The current experiments compare performance on a generalisation task after training using one of a number of different conditions including Free Classification, allowing a comparison with participants who have had a teaching signal during training. The results

. Conclusions: A system that combines the output of an NLP system with the standard techniques of text categorisation can produce results that exceed the performance of either system on its own. The F1 of a system that combined features of an NLP system with standard text categorisation features was 68

Abstract. This paper presents an extension of prior work by Michael D. Lee on psychologically plausible text categorisation. Our approach utilises Lee’s model as a pre-processing filter to generate a dense representation for a given text document (a document profile) and passes