在具有模式的完整统计知识条件下，按照贝叶斯决策理论进行设计的一种最优分类器。分类器是对每一个输入模式赋予一个类别名称的软件或硬件装置，而贝叶斯分类器是各种分类器中分类错误概率最小或者在预先给定代价的情况下平均风险最小的分类器。-In a model under the condition of complete statistical knowledge, in accordance with the Bayesian decision theory to design an optimal classifier. Classifier is the input mode of each name given to a category of software or hardware devices, and Bayesian classifier is the classifier in the smallest classification error probability or the cost of pre-given circumstances smallest average risk classifier.