期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2013
卷号：10
期号：6
出版社：IJCSI Press
摘要：This paper aims at developing a system that is capable of classifying Arabic and English un-structured documents; it proposes to classify these documents in consecutive two phases. In the first phase, incremental Automated Domain-Meta-Document Construction (ADC) algorithm is applied as a new automated machine learning approach. ADC constructs updatable summarized Domain-Meta-Documents, which corresponds to the trained classified documents. The output would be stored in a knowledge base in order to help in the classification process. In the second phase, an enhanced supervised classification algorithm based on automated calculation of threshold value would utilize the previously generated Domain-Meta-Documents to classify the incoming Dataset. To evaluate the performance of this proposed approach, two experiments were conducted on two standard dataset, namely Corpus of Contemporary Arabic (CCA) and Newsgroup 20, whose results revealed that the proposed classification approach outperformed the compared classification algorithms (C4.5 and Back Propagation Neural Network) in different measures. The general accuracy of the proposed system was found to be about 95%.
