Multi-class text categorization based on LDA and SVM
Kunlun Li, Jing Xie, Xue Gang Sun, Yinghui Ma, Hui Bai · Procedia Engineering · 2011
When dealing with the high dimensions and large-scale multi-class textual data, it is commonly to ignore the semantic relation between words with the traditional feature selection method. In order to solve the problem, we introduce the categories information into the existing LDA model feature selection algorithm, and construct SVM multi-class classifier on the implicit topic-text matrix. Experimental results show that this method can improve classification accuracy and the dimensionality is reduced availably, the value of F1, Macro-F1, and Micro-F1 are obtained improvement.