Textual topic evolution analysis based on term co-occurrence: A case study on the government work report of the State Council (1954–2017)

The government work report of the State Council is a kind of comprehensive policy text. This paper uses text mining technology to carry out a comprehensive multi-granularity, multi-level quantitative analysis of the government work reports, which has a great practical and instructive significance for relevant personnels to understand the evolution of domain knowledge in a short time. Firstly, a series of text preprocessing is done by using the Chinese word segmentation tool combined with three kind of dictionary built by authors, i.e., the domain word dictionary, the domain synonym dictionary and the domain stopword dictionary. Then, according to the co-occurrence information of words in the government work reports, we attempt to conduct topic modeling on the corpus consisted of all the government work reports and single government work report respectively, Finally, we find 12 latent topics for the corpus, such as the "Economic reform", "Agriculture", "Government construction", "Defense military" and so on. Based on the 12 topics, we conduct the topic modeling on every single government work report, with which topic evolution analysis is carried out over the whole period of all government work reports.