Topic Discovery in Research Literature Based on Non-negative Matrix Factorization and Testor Theory

The paper proposes a new way of comprising the Non-negative matrix factorization (NMF) and Testor theory to make topic discovery. NMF method is good at dealing with high dimensional documents and clustering, while Testor theory is used to find the topic of each cluster. By an example of ten abstracts of Chinese science literature from magazines relative to environmental science, the whole process is described in detail. In the end, a case study about automatic classification of a conference proceeding (in Chinese) is given. The result shows the effectiveness of the whole method.