Automatic Determination of Semantic Attributes for User Defined Words in Japanese to English Machine Translation,

*Satoru Ikehara, *Satoshi Shirai, *Akio Yokoo   *Francis Bond and **Yosie Omi

*NTT Communication Science Laboratories
**NTT Advanced Technology Company

Abstract

This paper proposes an automatic determination method of semantic attributes for user defined words. This method determines semantic attributes of the user defined words, both Japanese index word and the English translation of which are given by a user.

Applying this method to compiling of a user dictionary required for newspaper article sentences and sentences of software design documents has revealed the following. More semantic attributes are rendered by the proposed method in comparison to the Manual Method and between 50 to 80% recall factor were achieved. Further, the translation quality is on a level equivalent with that resulting from use of a manually compiled user dictionary. From the foregoing,



[ In Proceedings of Workshhop, 7th Annual Meeting of ANLP, pp.3-8 (July, 1994). ]