Homework 3
• 32 min read • 6268 words
Tags: Ma-Le Probability
Categories: Machine Learning
Homework 3
2. Gaussian Classification
Let for a two-class, one-dimensional () classification problem with classes and , , and .
Find the Bayes optimal decision boundary and the corresponding Bayes decision rule by finding the point(s) at which the posterior probabilities are equal. Use the 0-1 loss function.
对于0-1损失函数,贝叶斯最优决策规则是:对于一个给定的观测值 ,我们应该选择后验概率 最大的那个类别。而决策边界即为两者的后验概率相等:
也即:
由于两个类别的先验概率相同,有:
带入正态分布PDF中:
决策边界为两个均值的中点。
Suppose the decision boundary for your classifier is . The Bayes error is the probability of misclassification, namely
Show that the Bayes error associated with this decision rule, in terms of , is
将先验概率带入这个公式:
决策边界是 并且 ,自然的分类规则是:
的实际概率为,这可以通过下面的积分得到:
的实际概率为,这可以通过下面的积分得到:
带入 ,整理即得:
Total words: 6268
Comments