案例推理从20世纪90年代初在人工智能领域兴起,在各个领域得到了广泛的发展。由于地理现象的复杂特性,许多现象发生的本质或机理还有待探索,其定量化的模拟和预测还不成熟,因此应用案例推理成为地学数据定量分析的新方法。但地理案例自身特殊的时空特征决定了地理案例的推理方式不同于传统的方式,需要考虑案例自身或者案例之间时空上的相互依赖关系。本文在分析地理案例时空分布特征的基础上,给出了案例推理的地学推理模式。 Case-based reasoning has arisen in the field of artificial intelligence since the early 1990s and has been widely developed in various fields. Due to the complex nature of geographical phenomena, the nature or mechanism of many phenomena still remains to be explored. The quantitative simulation and prediction are not yet mature. Therefore, case-based reasoning has become a new method for quantitative analysis of geoscience data. However, the special spatiotemporal features of geographical cases determine that geographical cases are reasoned differently from the traditional methods and need to consider the temporal and spatial interdependence between the cases themselves or the cases. Based on the analysis of the temporal and spatial distribution of geographical cases, this paper presents a geo-logical reasoning model for case-based reasoning.