We propose a method to build thesauri on the basis of grammatical relations. The proposed method constructs thesauri by using a hierarchical clustering algorithm. An important point in this paper is the claim that thesauri in order to be efficient need to take (surface) case information into account. We refer to the thesauri as "relation-based thesaurus (RBT)." In the experiment, four RBTs of Japanese nouns were constructed from 26,023 verb-noun co-occurrences, and each RBT was evaluated by objective criteria. The experiment has shown that the RBTs have better properties for selectional restriction of case frames than conventional ones. 1 Introduction For most natural language processing (NLP) systems, thesauri are one of the basic ingredients. In particular, coupled with case frames, they are useful to guide correct analysis [ Allen, 1988 ] . In the example-based frameworks, thesauri are also used to compensate for insufficient example data [ Sato and Nagao, 1990, Nagao and Kurohashi...

In example-based NLP, the problem of computational cost of example retrieval is severe, since the retrieval time increases in proportion to the number of examples in the database. This paper proposes a novel example retrieval method for avoiding full retrieval of examples. The proposed method has the following three features, 1) it generates retrieval queries from similarities, 2) efficient example retrieval through the tree structure of a thesaurus, 3) binary search along subsumption ordering of retrieval queries. Example retrieval time drastically decreases with the method. 1 Introduction Since a model of machine translation (MT) called Translation by Analogy was first proposed in Nagao (1984), much work has been undertaken in examplebased NLP (e.g. Sato and Nagao (1990) and Kurohashi and Nagao (1993)). The basic idea of examplebased approach to NLP is to accomplish some task in NLP by imitating a similar previous example, instead of using rules written by human writers. Major pro...

The nearest neighbor algorithm is the most basic class of techniques in the sub-fields of machine learning such as case-based reasoning (CBR), memory-based reasoning (MBR), and instance-based learning (IBL). In the nearest neighbor algorithm, the computational cost of example retrieval is one of the most important issues. This paper proposes a novel technique for optimizing the nearest neighbor algorithm. Its basic idea is based on the use of similarity template, which is a data structure that enumerates all the possible patterns of calculating similarity between two examples. In the method, the nearest neighbor retrieval process is optimized by generating retrieval queries from an input and similarity templates in a certain order. Its major advantages are as follows: 1) unlike the hierarchical organization of memory, it is guaranteed that examples with the greatest similarity value according to the given similarity measure are retrieved, 2) it is easy to add new examples to the examp...

This paper presents the KN parser, a Japanese dependency/case structure analyzer. It performs well in parsing Japanese sentences by taking account of certain characteristics in Japanese language. Several unique methods used in the parser are discussed and preliminary evaluation results on hundreds of Japanese sentences are given. 1 Introduction This paper presents the KN Parser (KurohashiNagao Parser), a dependency grammar parser that takes advantage of several unique methods that have been recently developed [1]. Figure 1 shows the overall flow of the system when parsing a sentence. One of the unique and most effective parts of the parser is its first component, namely, the coordinate structure analysis component. This component drastically reduces structural ambiguity in a sentence containing coordinate structures, making the subsequent analysis much easier. Basically the KN parser works as a dependency structure analyzer based on dependency grammar formalism. Using a wide-coverage s...

In example-based NLP, the problem of cmnputationM cost of example retrieval is severe, since the retrieval time increases in proportion to the number of examples in the database. This paper proposes a novel example retrieval nethod for avoiding full retrieval of examples. The proposed method has the following three features, ) it generates retrieval queries from similarities, 2) ef- ficient example retrieval through the tree structure of a thesaurus, 3) binary search along subsumption ordering of retrieval queries. Exmnple retrieval time drastically decreases with the method.