Identifying the stages of a technology's lifecycle is crucial for predicting technological trends, formulating science and technology policies, and corporate strategies. This study proposes a dynamic network detection-based method for technology lifecycle identification, integrating dynamic citation networks and co-classification networks. By quantifying network metrics and applying segmented fitting models, this approach delineates lifecycle stages and comprehensively assesses technological development trends. Using thin-film transistor liquid crystal displays (TFT-LCD) and nanobiosensors (NBS) as examples, the feasibility and effectiveness of the proposed method are validated, along with rules for dynamic network-based lifecycle segmentation, providing reference for technology lifecycle identification. Findings indicate that while network segmentation results for TFT-LCD and NBS technologies exhibit temporal discrepancies from actual lifecycle stages, they maintain high consistency. Additionally, the dynamic network detection method features a 2-4 year lag window, attributed to patent disclosure delays. Compared to traditional S-curve and multi-indicator methods, this approach demonstrates higher robustness and precision, reduces reliance on data quality, and effectively reveals technological development phases, offering a new theoretical framework and practical reference for technology lifecycle identification and prediction.
评论