论贝叶斯学习机的学习过程

贝叶斯学习是机器学习研究的一个重要方向,它是以贝叶斯定理为基础,基于已知的概率分布和观察到的数据,并结合先验知识进行推理,作出最优决策的一种概率手段. 本文首先针对参数和变量的不同类型分别给出四种情形的贝叶斯公式,然后结合一个指数分布的特例,研究了贝叶斯学习过程中有关信息的转换过程,指出了如何合理正确地利用先验信息、模型信息和样本信息.

The Learning Mechanism of Bayesian Learning Machine

Bayesian learning is an important research direction in machine learning,which is a probability method of making optimal decision based on prior knowledge,given probability distributions and observed data.Bayesian Theorem is the foundation of this process.Four kinds of Bayesian formulae for different types of parameters and variables are proposed in this paper respectively.The mechanism of information conversion in Bayesian learning is explored to illustrate how to combine the prior,population and sampling ...