基于时间概率分布的时序知识图谱逻辑推理方法及装置

• Format: Patent
• Language: Chinese
• Published: 01.11.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

本申请提出了一种基于时间概率分布的时序知识图谱逻辑推理方法及装置，涉及大数据、人工智能与知识图谱技术领域，其中，时序知识图谱基于文本数据生成，文本数据中的每段文本表示现实世界的一个事件，时序知识图谱以#imgabs0#的形式存储现实世界的一个事件，该方法包括：通过链接预测器补全时序知识图谱；将补全的时序知识图谱建模为时序概率分布，并将时序概率分布通过筛选转化为稀疏矩阵形式；针对用户输入的逻辑查询，在稀疏矩阵形式的时序概率分布上进行模糊逻辑运算，得到从起点到终点的中间路径以及对应的概率分布，并将概率值最高的路径作为查询结果。采用上述方案的本申请能够有效提高对复杂时间逻辑查询的处理能力。 The invention provides a time sequence mapping knowledge domain logical reasoning method and device based on time probability distribution, and relates to the technical field of big data, artificial intelligence and mapping knowledge domains, a time sequence mapping knowledge domain is generated based on text data, and each text in the text data represents an event in the real world. The time sequence knowledge graph stores an event in the real world in the form of # imgabs0 #. The method comprises the following steps: complementing the time sequence knowledge graph through a link predictor; modeling the complemented time sequence knowledge graph as time sequence probability distribution, and converting the time sequen