模式识别Chapter 2-1.ppt

基于最小风险的贝叶斯决策 * Likelihood ratio : Decision rule: If the likelihood ratio of class and exceeds a threshold value (that is independent of the input pattern ), the optimal action is to decide . * Decision functions A feature vector x is assigned to class if 分类风险最小 Minimum-error rate classification * 正确类别的后验概率=正确率 True state : , correct rate= 错误率=1 - 正确率 Minimum-error rate classification * Minimum-error rate classification * error rate Minimum-error rate classification * * Bayes’ Decision: error minimum The probability of misclassification is minimized by selecting the class having the largest posterior probability * Bayes’ Decision: risk minimum If the likelihood ratio of class and exceeds a threshold value (that is independent of the input pattern ), the optimal action is to decide . * 小结 分类问题： 数学问题： 如何解决这个数学问题？ 数学问题 从输入到输出的映射 概率、 Bayes 决策 最小错误率：后验概率 最小风险：似然比 * 作业 二 1. 两类样本C1， C2的特征分布直方图如下：试求每类样本的先验概率、(类)条件概率及后验概率。 * 作业 二 2. 已知甲类：P(ω1) = 0.7和类条件概率密度函数p(x|ω1) ,乙类：P(ω2) = 0.3和类条件概率密度函数p(x|ω2) 今有待分类样本特征观察值x = 10，且由函数曲线查得p(10|ω1) = 0.2， p(10|ω2) = 0.5 ⑴试用最小错误率Bayes决策对样本x = 10进行分类 ⑵试用最小风险Bayes决策对该样本进行分类，设λ11=λ22=0，λ12=2，λ21=1 * Bayesian Decision Theory Bayesian decision theory is a fundamental statistical approach to the problem of pattern classification. The approach is based on quantifying the tradeoffs between various classifications. It makes the assumption that the decision problem is posed in probabilistic terms, and that all of the relevant probability values are known. * Bayes’ theorem Conditional probability -- specifies the probability that the observation falls in column of the array given that it belongs to the class 条件概率 -- 类 ，同时特征值为 * Bayes’ theorem * Bayes’ theorem * Bayes’ theorem * Bayes’ theorem * Bayes’ theorem What is The probability that the class is given that the measured value falls in the cell Posterior