Research on the construction of computer time series knowledge graph based on the transformation of knowledge dissemination paradigm

The paradigm shift of knowledge dissemination has accelerated the speed of knowledge generation and dissemination. Traditional models are difficult to handle massive, rapidly updated and complex knowledge flows, and cannot meet the needs of various fields for real-time and effective knowledge. Although computer time-series knowledge graphs can present the evolution of knowledge over time and the dynamic relationship between entities, there are problems in the construction of traditional algorithms, such as difficulty in adapting to the dynamic changes of knowledge dissemination and shallow semantic understanding. This study proposes a time-series knowledge fusion algorithm (DTWSK) based on dynamic time windows and semantic associations. Through the dynamic time window mechanism, the time range is adjusted in real time to accurately capture the knowledge changes at key time nodes; using deep learning and natural language processing technology, the complex semantic associations between knowledge are deeply mined from the dual dimensions of knowledge graph structure and text semantics, and a more accurate, comprehensive and real-time time-series knowledge graph is constructed. Experimental verification shows that Compared with traditional algorithms, the DTWSK algorithm demonstrates superior performance with accuracy (85%), recall (82%), and F1-score (83%), particularly in complex semantic and dynamic knowledge dissemination scenarios such as tracking disease progression in the medical field. It has obvious advantages, effectively improving the performance of temporal knowledge graph construction and promoting the development of knowledge services towards intelligence and refinement.