Thesaurus-Based Semantic Compression for Afaan Oromo Text Retrieval

In this study, we have developed AOIR using thesaurus based semantic compression approach. The developed prototype has both indexing and searching parts. For an experiment, 106 Afaan Oromo text documents collected from news articles, the Afaan Oromo Bible, websites and books by the researchers. Then, different text operations were used to identify content-bearing terms. Finally, semantic compression is applied in which the semantic representation of each term is identified and thesaurus based inverted indexing structure is used in document indexing. The experiment is conducted using two ways, the first experiment is conducted without considering the synonymous search term and the second experiment is conducted by using thesaurus based indexing. From conducted experiment, the average result of 58.451% precision, 78.438% recall, and 61.520% F-measure are registered without thesaurus based semantic compression approach and thesaurus based semantic compression technique showed 71.014%, 89.287% and 77.715% average precision, recall, and F-measure respectively.