中文摘要：

归纳推理是指从个别前提推出一般结论的推理，它是人的一项基本认知能力，也是心理学研究需要解释的重要现象。在模拟归纳推理时，不能排除先验知识对归纳推理结论力度的影响，贝叶斯理论可以结合条件对先验概率进行修正，能够很好地将先验知识引入到归纳推理中来。文章重点介绍了基于记忆的贝叶斯模型，相似性概率模型和结构统计模型对归纳推理的解释。

英文摘要：

Inductive reasoning can make powerful generalizations from sparse data. It is a basic capacity and a phenomenon that psychol- ogist should explain. The background knowledge plays an important role in inductive reasoning. So the computational models of inductive reasoning will be imperfect without knowledge. Bayesian theory can revise prior probability by combining the evidence. Bayesian infer- ence in probabilistic models can explain how prior knowledge is used in inductive reasoning. The paper introduces three probabilistic models of inductive reasoning： memory - based Bayesian model, SimProb model and structured statistical mode.