Browse Prior Art Database

Simplified dictionary structure for morphological analysis Disclosure Number: IPCOM000018810D
Original Publication Date: 2003-Aug-13
Included in the Prior Art Database: 2003-Aug-13
Document File: 1 page(s) / 71K

Publishing Venue



A program is disclosed that an simplified dictionary structure for morphological analysis. The purpose of introducing this structure is to improve runtime performance of morphological analysis. Morphological analysis program generally uses at least two data to carry out processing. One is word dictionary, and the other is grammatical connectivity rules. Thus far, these two data was stored and referred in separate structures. That was not good from the time performance viewpoint. To solve this problem, it is effective to combine these two data into one simple structure at preprocess stage.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Simplified dictionary structure for morphological analysis

  Each entry in word dictionary contains a pair of information, word's surface string and its associated POS (Part of Speech) code. Grammatical connectivity rules define the adjacency relations between each words. Those relations are determined according to POS codes. Morphological analysis program uses these 2 data as following steps.

Lookup word dictionary, and get its associated POS code.

According to POS code, search through grammatical connectivity rules to get

corresponding definitions. Using derived definitions, solve ambiguity on word boundary and POS codes in input




  sentence. However, both data in word dictionary and grammatical connectivity rules are settled entirely before starting analysis processing. So if these two data were combined in one structure at preprocess stage, it would be able to achieve performance improvement on analysis runtime. Since above step1 and step2 would become a single step. Following figure shows this combining preprocess.

Disclosed by International Business Machines Corporation


[This page contains 55 pictures or other non-text objects]