中文摘要：

为了扩展自动机的学习范围,提出了在理论上将加权自动机的学习能力从域扩展到有限半环的方法,在这个扩展的基础上,提出了一种基于精确学习模型的学习算法.针对该算法分析得出并不高于已有学习算法的时间复杂度,基于扩展了的定义域给出可以学习的实例,通过实验给出了一套对比学习确定加权自动机算法的实验结果,对状态缩减程度进行了说明和分析,阐述了该算法的正确完备性和更广的适用范围.

英文摘要：

To expand the scope of learning automata,a new method which extended domain of learning weighted automata from field to semiring was proposed,on the basis of this expansion,a new algorithm based on the exact learning model was proposed.The new algorithm had no higher time complexity than other learning algorithms and could learn examples on the new domain.The experimental results on a collection of examples confirmed the space advantage of learning algorithm for deterministic and weighted automata,and finally demonstrated the new algorithm was correct,complete and had broader scope.