There is limited research on question answering based on temporal knowledge graphs,with most approaches being confined to semantic or temporal information matching,and lacking the utilization of temporal constraint information.A method for question answering based on temporal constraint subgraph extraction is proposed in this paper,whereby subgraphs are pruned based on time range information,and subsequently,answer inference is performed utilizing the pruned subgraph and temporal information to derive the final answer.Experimental results demonstrate that the proposed method,compared to baseline models,exhibits significant improvements in answering complex and temporal-type questions.
维普
万方数据