Improving Afaan Oromoo Information Retrieval Using Query Expansion

Abstract

In This Era Technology Has A Great Role In The Development Of One Language And Accessing The Right Information Written In A Different Language Is Necessary Which Can Be Considered As One Of The Basic Human Needs. So Designing An Effective Performance Of Information Retrieval System Is Important To Search And Retrieve The Most Relevant Document From Online Or Offline. As Afaan Oromoo, All Relevant Documents Cannot Be Retrieved Because Of Ambiguous Words And Relevance Ranking Problems That Don't Satisfy Users?�? Information Needs. Afaan Oromoo IR System Is Developed To Improve Searching And Retrieving Of Information Written In Afaan Oromooand Make This Language Proceed With Current Technology That Satisfies Users Of This Language. In This Study,We Developed An Information Retrieval System Using One Of The Probabilistic Model Approach Called BM25 By Integrating With An Elastic Search That Makes The Searching Process Fast. However, BM25 Improves The Search Result By Optimizing The TF-IDF Score For Query, Word Ambiguity Decreases The Effective Performance Of Afaan Oromoo Information Retrieval As One Word Can Have Multiple Meanings. Thus, The Query Expansion Technique Is Implemented To Improve The Retrieval System By Disambiguating Those Words. Afaan Oromoo Wordnet Is Developed With Synsets And Gloss Method Construction To Reformulate The Query And The Query Is Modified Using Semantic Relation Through Synsets To Synsets Method. Finally, The Query Expansion With BM25 Is Integrated With Afaan Oromoo Information Retrieval Systems To Enhance And Improve Retrieval Performance. As Per The Experimental Results Carried Out, The System Registers Different Values Before And After Query Expansion In Terms Of Recall, Precision, And F-Measure That Is The System Registered 87.15%, 82%, And 83.3% Respectively Before Query Expansion, And After The Query Is Expanded The Retrieval System Registered 95.9%, 84.2%, And 89.2% Respectively. Generally, The Retrieval System Is Improved By A 5.99% F- Measure From The Original Query. Therefore, Developing Query Expansion With BM25 Records Better Results And Improves The Performance Of The Afaan Oromoo Retrieval System. However, The Retrieval System Performance Might Be Improved Better If Other Semantic Relations Are Included In Future Work.