Text mining based on the self-organizing map method for Arabic-English documents.
Abdulsamad Al-Marghilani, Hussein Zedan, Aladdin Ayesh · DMU Open Research Archive (De Montfort University) · 2008
Computer information and retrieval is becoming increasingly sophisticated and is being exploited in more and more spheres of human activity. Many computer applications are developed as information distribution systems, of which the Internet is one of the best known and widely used. With enormous quantities of data in different languages available on the net, it is essential that more efficient methods of language data extraction are daveloped. Thus this paper is focused on text mining multilingual datasets. Arabic is a highly derivated and inflected language, requiring proper morphological analysis for effective text mining, and yet no standard approach to word stemming has emerged. This work is an attempt towards the development of a tool useful in the analysis of Arabic-English texts, and is achieved through the multilingual text mining (MTM) of a combined Arabic-English corpus. This project is based on SelfOrganizing Map (SOM) and uses an Arabic-English text corpus as the test-bed. Issues related to Arabic-English text mining, stemming and clustering are discussed in this paper. To the author’s knowledge, there is no significant literature available regarding SOM techniques applied to Arabic-English language text mining. In this work a framework and the outcome of its implementation is presented.