Design and Implementation of IR System for Tigrigna Textual Documents

Abstract

Nowadays, various amount of information's are available on the internet. To search relevant documents from the internet development of information retrieval system or search engines is necessary. Therefore, this paper deals with development of Information Retrieval system for Tigrigna textual documents. It helps to find relevant documents from the internet, which are stored in Tigrigna language for the Tigrigna language users to satisfy their information need. The system includes two sub systems those are indexing and searching part. The indexing part is the process of organizing filtered Tigrigna documents using keywords extracted from the entire Tigrigna collection or corpus. It is an offline process carried out by the producers or authors world to speed up searching of information from the entire document as per users query. Searching is the process of scanning documents to find relevant documents that matches to the users query or information need. It is an online process mostly carried out by the users or readers world. Vector space model techniques was applied to implement this system. Vector space model is the most core information retrieval technique used to calculate similarity measure between the query and the documents finally it ranks the most relevant documents to the given query according their similarity score in descending order. According to this, the retrieval system was tested and the experimental results of the system in Tigrinya documents returned an encouraging and promising result. The system has registered, 70% precision and 84% Recall.