Topic Mining and Effectiveness Evaluation of China’s Coal-Related Policy Based on LDA Model

Policy text mining is widely used to evaluate the evolution of policy themes and the effects of policy interventions. In this paper, focusing on China’s most abundant energy resource, latent Dirichlet allocation (LDA) model is adopted to recognize the technical themes included in the coal-related policies as well as the emphasis of policies on technical themes. Combined with publication years of the policies, the yearly strengths of the themes are calculated for evolution analysis of the technical themes. To evaluate the effects of policy interventions, technical themes included in the coal-related projects funded by the National Natural Science Foundation of China (NSFC), and their evolutions are also analyzed and cross-certificated with those acquired from policies. The analysis shows that the evolutions of policy themes agree with the nation’s vision of energy transition, and the technical themes of the NSFC projects evolve in a similar trend as those of policies, with a delay of 5–6 years.