Arabic information retrieval system using the neural network model

Abstract: Information Retrieval (IR) for Arabic language has gained significant attention and emerged as one of the research topics that has been studied by Arabic and foreign researchers. The goal of this research is to apply the IR using Neural Network (NN) Model on natural Arabic language text documents to solve the problem of retrieving the Arabic information from documents' database. Furthermore, all stored documents must be indexed with keywords classification that describe the exact content of each document, which makes it impossible to retrieve all related documents more computational time to classify and update the stored documents. IR using NN applies to solve the problem of documents indexing, classification and retrieving the related documents using Terms of weight and Normalization. The computational results have been compared with the Vector Space Model (VSM) and showed an improvement of NN training time compared with VSM load document time.

Keywords: Arabic, Information Retrieval, Neural Network, Vector Space.