Effective feature selection is essential to make the learning task efficient and more accurate. In this paper, a novel Chinese text feature selection algorithm based on PLSA was presented for text classification, and it was compared with other effective feature selection methods on a benchmark of 8 text classification problem instances that were gathered from Sougou Lab's corpus. The results show that this method could make SVM classifier have the best classification performance and more robust than other methods.
展开▼