Abstract: Markov model plays an important role in the recognition field of standard text and hand-writing．In order to consider how to reestimate model-parameters from observation data，the criterion function introduced by the entropic propcrtles of the model replaces Baum auxiliary function，and formulas of reestimating model-parameters is obtained using the properties of the criterion function introduced by the entropic properties ofthemode1．

Key words: Mmkov model, entropic, criterion function, reestimating model-parameters

摘要： Markov模型在标准文本和手写文字的自动识别领域中有特别重要的应用．在讨论如何由观测数据来估计模型参数的问题中，代替Baum辅助函数采用模型的熵特性引入判断函数，并利用此判断函数获得模型参数的重估算法．

关键词: Matkov模型, 熵, 判断函数, 参数重估