Modelling and Designing of NLP Interface to Database for Afaan Oromoo

Abstract

Database management system has been  widely used for storing and retrieving data. However,  database is often hard to access the data since their interface  is rigid in cooperating with user, due to that analysis of  natural language query interface to relational databases has  gained much interest in research community. This can be  termed as structured free query interface as it allows the  users to retrieve the data from the database without knowing  the underlying schema. Structured free query interface  should address majorly two problems. Querying the system  with natural language interfaces is comfortable for the naive  users but it is difficult for the machine to understand. The  other problem is that the users can query the system with  different expressions to retrieve the same information. The  different words used in the query can have same meaning  and the same word can have multiple meanings. Hence; it is  the responsibilities of the NLI to understand the exact  meaning of the word in the particular context to handle such  problem the researcher develops an algorithm that is from  manually crafted rule that extracted from user queries for  proper query translation and SQL query generation. The  proposed, pattern matching approach to natural language  interface to database system has been passes through  various phases. The idea behind the approach is to  overcome the problem bottleneck luck of tools in the  language, while pattern-matching method can fill the gap of  those resources. This makes our approach suitable for  AONIDB since Afaan Oromoo is under resourced  language. The exact meaning of the word used in the query  in particular context obtained using domain-mapping  dictionary and database-mapping table constructed form  student database that contains grade, department and course  table for mapping database information with the help of an  algorithm developed by researcher based on the rule that  extracted from collected user query. The system has been  developed in which user can input query in Afaan Oromoo  language and can see result in the same language. The  proposed system can handle list query, single and multiple  conditional queries, aggregate function queries and join  queries. The proposed system evaluated on 100 Afaan  Oromoo queries using student database with precision,  recall and F-measure. The result of testing is very  encouraging and an average performance of 94.2% that  implies that the systems validity and reliability is very high,  an indicator of its strong and successfully features use and  operation. Yet; further experiments using different  approaches that extend this work are needed for a better  performance.