Cornelis H A Koster

Learn More
Natural language processing techniques are becoming more and more popular in improving the eeectiveness of retrieval and ltering systems. These systems usually incorporate a kind of natural language parser based on a grammar which describes the natural language or a part of it. In the context of the information ltering project Profile, a framework for the(More)
A normal table of events of the first cleavage period in the fertilized egg (cf. Gerhart, 1980) has been completed (cf. Table I) by studying external and internal features. Through a cytological study of eggs fixed after video time-lapse observation such features can directly be correlated and it has been shown that the first postfertilization wave (PFW)(More)
A previous version of this work was presented as a paper at RIAO 97 conference in Quebec, Canada and included in RIAO 97 proceedings (Arampatzis et al., 1997a). Abstract In this article we describe a retrieval schema which goes beyond the classical information retrieval keyword hypothesis and takes into account also linguistic variation. Guided by the(More)
AAx Grammars over a Finite Lattice (agfls), a simple form of two-level grammars admitting quite eecient implementations, are proposed as a formalism to express the syntax of natural languages. In this paper the concepts and notation of agfls are described. A brief example is given of their use in describing a fragment of the English language, followed by a(More)
In this paper a conceptual model for object-oriented analysis is introduced. Three submod-els are described which can be seen as milestones during the analysis phase. Each (sub)model has a corresponding paraphrasing mechanism which may be used (1) to provide a description of the structure of the model and (2) to generate sample instantiations. This(More)
AAx Grammars over a Finite Lattice (agfls), a simple form of two-level grammars admitting quite eecient implementations, are proposed as a formalism to express the syntax of natural languages. In this paper the concepts and notation of agfls are described. A brief example is given of their use in describing a fragment of the English language, followed by a(More)
We present the differences beetween the results of two series of experiment of supervised document classification with Winnow: first on a selection of typical documents in a small number of classes; then on all the documents in all the classes. We see that the optimal parameter values for the subset are not optimal for the whole set. And the results on the(More)