Purpose The purpose of this study is to address the challenges of ambiguous topics and unclear development paths in technological evolution by identifying disruptive technology topics. This work is of strategic significance for nations to proactively plan for and lay out future industries. Design/methodology/approach This paper innovatively integrates natural language processing techniques with the latent Dirichlet allocation (LDA) topic model to construct a disruptive technology topic identification framework based on multi-source heterogeneous data. Taking the field of quantum communication as a case study, it selects scholarly papers and patents as heterogeneous data sources and uses multi-dimensional indicators – including topic similarity, novelty and intensity – to identify technology topics. Findings The results indicate the following: the complementary fusion of scholarly papers and patent data significantly enhances the comprehensiveness and accuracy of technology theme identification. The classification framework based on theme novelty and intensity effectively distinguishes technical characteristics at different development stages, providing a quantitative basis for optimizing innovation resource allocation. The algorithm-driven theme identification method offers an extensible analytical tool for technological foresight. Research limitations/implications Future research on identifying disruptive technological innovations should focus on the dynamic weighted fusion of multi-source data, the precise management of technology life cycles and the balance between quantitative analysis and situational flexibility. Practical implications This study provides scientific decision support for the research and development of disruptive technologies. Originality/value This study lays a methodological foundation for the strategic layout of future industries by proposing a novel, data-driven framework for technology foresight.
Liu et al. (Thu,) studied this question.