Modeling and Designing Amharic Query System to Bilingual (English-Amharic) Databases
No Thumbnail Available
Date
2014-04
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
for information is a very important part of the day to day activities of human beings. Information can be acquired from different sources, one of the major sources being databases. Formulating a query in a way the computer system understands is necessary to retrieve data from a database. To accomplish that, one should write SQL statements. But everyone is not able to write SQL queries because of its requirement to be able to know the technical structures of SQL query formulation. Instead of formulating SQL statements, using natural language to retrieve data from database prompted the development of Natural Language Interface to Database Systems (NLIDB). NLIDB allows translating Natural language queries into SQL queries. Hence more friendly to naïve and non technical information seekers to use their own language to access database contents.
Research on NLIDB started in the late sixties. Since then, many research works have been conducted. A lot of researches have been done for different languages like English, Arabic, Spanish, Punjab, etc. However, little is done to design NLIDB in Amharic language. Only one work is available that attempted to model and design Amharic Structural Query Language (ASQL). The model was syntax based. It requires users to know the syntax of the ASQL to formulate their queries. In this work, we designed a NLIDB system for Amharic. With this Amharic Interface users can extract information from database systems using plain Amharic statements without being constrained to query language syntax.
The Amharic Interface to databases translates the plain Amharic statements into English SQL queries. We designed a general architecture for Amharic NLIDB; we designed different algorithms to extract database elements from databases, to manage Amharic statement pre-processing, to map the translated tokens to database elements, and to generate equivalent SQL statements. Furthermore, we have identified and used the necessary tools and resources such as tokenizer, normalizer, stop word remover, Amharic stemmer, bi-lingual (Amharic-English) translation dictionary, list of thesaurus, list of stop words, etc. Finally we have implemented and tested the Amharic Interface for Database query. The evaluation of the system shows encouraging results.
Keywords: Amharic Query Interface for Databases, Amharic Query System for Bi-lingual databases, Query Translator, Natural Language Query Mapper, Natural Language Interface for Databases (NLIDB), Database Element Extractor.
Description
Keywords
Amharic Query Interface for Databases, Amharic Query System for Bi-Lingual Databases, Query Translator, Natural Language Query Mapper, Natural Language Interface for Databases (NLIDB), Database Element Extractor