Abstract: 【Purpose/significance】Disruptive technologies are concealed and mutated, and technological topics will release early sig?
nals of disruptive trends during the evolution process. This paper proposes a method to identify disruptive technology based on the evo? lution of patent themes.【Method/process】Firstly, the technology life cycle is judged based on the S-curve method; Secondly, methods on text mining and social network analysis are adopted to identify the evolutional events of technological topics; finally, three types of mutation situations are defined aim at selecting technological topics with high disruptive potential.【Result/conclusion】Empirical study on patents of brain-inspired intelligence has discovered six emerging topics, four hot ting topics and six merging topics. Through expert interpretation, the effectiveness of the proposed method is proved, and five topics are screened out as the final prediction re? sults.【Innovation/limitation】This paper analyzes the evolutionary characteristics of technical topics based on the dynamic community discovery algorithm and then screens disruptive technologies based on topic mutation. However, the data source is relatively single and will be expanded in future research.