An Integrated Development of a Query-Based Document Summarization for Afaan Oromo Using Morphological Analysis

Abstract

Objective: To develop document summarization for the Afaan Oromo language based on the query entered by the user(s). Methods: This study follows the design science analysis technique as a result of its considerations of thoughtful, intellectual, and ingenious activity throughout problem-solving and the creation of knowledge. The developed query-based framework has used the TF-IDF term weight methodology. Development tools such as HornMorpho are employed for morphological analysis; whereas, Natural Language Processing Toolkit is employed for the text process. The system has experimented on the various extraction rates of 10%, 20%, and 30%. The result’s evaluated exploitation recall, precision, and F-measure for objective analysis; whereas, subjective analysis has been evaluated by language consultants. Findings: The results of the evaluations showed that the proposed system registered f-measure of 90%, 91% and 93% at a summary extraction rate of 10%, 20%, and 30% respectively. The informativeness and coherence of the proposed system also registered its best performance summary of 51.67%, 56.67 % and 54.17% average score on five scale measures at an extraction rate of 10%, 20%, and 30% respectively when both methods were used together. Novelty: By using a morphological analysis tool the performance of the system is improved from 80.67% to 91.3% F-measure when we compare it with the previous work even supposing there’s still a requirement to conduct additional analysis to enhance the Afaan Oromo text summarization. Keywords: Document; Summary; Natural Language Processing; Morphological Analysis; Text Ranking