中文摘要：

[目的／意义]科技创新需要快速发现特定科技领域中关键知识衍生与演化的路径，探索未来的知识创新趋势，为此，有必要对知识演化路径进行动态可视化研究。[方法／过程]从主题关联的角度入手，以3D打印领域为例，基于LDA识别出科技创新主题并进行分阶段细化分析，探测主题集群内部与外部的关联强度，识别出主题不同生命周期的演化能力及其演化类型。[结果／结论]实验结果表明，该方法从主题关联的角度入手，构建了基于时间序列的知识演化路径，丰富了知识管理和信息计量的理论研究方法，在实践上则有助于探测科技创新知识。

英文摘要：

[ Purpose/significance ] The research unable to meet the needs of knowledge management, of static visualization on the knowledge evolutionary path has been knowledge innovation and technological development. [ Method/ process] From the perspective of topics relevance, taking the case of 3D printing field, this paper identifies the scientific and technological innovation topics based on the LDA model and makes detailed analysis in every phase, probes the internal and external association strength of the topic clusters, and identifies the evolution ability and evolution type of topics in different life cycle. [ Result/conclusion ] Experimental results show that the method can build the knowledge evolutionary path based on the time series from the perspective of topic relevance. It can enrich the research methods of knowledge management and information measurement and help detecting the technological innovation in the practice.