Text classification based on Rough Sets theory is an effective method for the automatic document classification problem. However, the computing multiple reducts is a problem in this method. When the number of training document is large, it takes much time and large memory for the computation. It is very hard to be applied in the real application system. In this paper, we propose an effective way of data partition, to solve the above problem. It reduces the computing time of generating reducts and maintains the classification accuracy. This paper describes our approach and experimental result.
展开▼
