Abstract:Carrying out quantitative evaluation of IC industrial policy is the premise of ensuring scientific formulation and effective implementation of industrial policy, and is also an important basis for industrial policy adjustment and improvement. To understand the evolution of China's IC industry policy focus and the current defects and deficiencies, this paper used text mining, social network analysis, and Latent Dirichlet Allocation (LDA) topic model to mine the content features, structure features, and topic features of China's IC industry policy. Combined with previous studies, the PMC-AE index model is established to quantitatively evaluate the central and local policies of China's IC industry under the background of the new development pattern, and finally, the optimization path of industrial policies is proposed. The results show that the policy texts at the national and provincial levels are all at level Ⅰ, and the policy texts at the municipal and district levels are concentrated at levels Ⅱ and Ⅲ, reflecting that the national and provincial policy texts pay more attention to the play of the planning leading role, while the policy texts at the municipal and district levels pay more attention to the formulation and implementation of a specific policy measure and program. In general, the integrated circuit industry policies at all levels are generally reasonable, but there are also problems such as lack of predictability and supervision, lack of coordination and linkage of policy objects, and lack of full use of demand-based policy tools. Therefore, improving the industrial policy system, extending the policy coverage, strengthening the coordination of policy objects, and improving the policy tool combination are important paths for China's integrated circuit industry policy transformation under the new development pattern.