Text associative classification approach for mining Arabic data set
Abdullah Saeed Ghareb, Abdul Razak Hamdan, Azuraliza Abu Bakar · 2012
Text classification problem receives a lot of research that are based on machine learning, statistical, and information retrieval techniques. In the last decade, the associative classification algorithms which depends on pure data mining techniques appears as an effective method for classification. In this paper, we examine associative classification approach on the Arabic language to mine knowledge from Arabic text data set. Two methods of classification using AC are applied in this study; these methods are single rule prediction and multiple rule prediction. The experimental results against different classes of Arabic data set show that multiple rule prediction method outperforms single rule prediction method with regards to their accuracy. In general, the associative classification approach is a suitable method to classify Arabic text data set, and is able to achieve a good classification performance in terms of classification time and classification accuracy.