自然语言处理中,经常要处理序列标注问题(分词、词性标注、组快分析等),为给定的观察序列标注标记序列。
令o和s分别代表观察序列和标记序列,
根据贝叶斯公式,
1 生成模型和判别模型的定义
对o和s进行统计建模,通常有两种方式:
(1)生成模型 (产生模型)
构建o和s的联合分布p(s,o)
(2)判别模型 (条件概率模型, 条件模型)
构建o和s的条件分布p(s|o)
2 判别模型和生成模型的对比
(1)训练时,二者优化准则不同
生成模型优化训练数据的联合分布概率;
判别模型优化训练数据的条件分布概率,判别模型与序列标记问题有较好的对应性。
(2)对于观察序列的处理不同
生成模型中,观察序列作为模型的一部分;
判别模型中,观察序列只作为条件,因此可以针对观察序列设计灵活的特征。
(3)训练复杂度不同
判别模型训练复杂度较高。
(4)是否支持无指导训练
生成模型支持无指导训练。
3 二者的本质区别是
discriminative model 估计的是条件概率分布(conditional distribution)p(class|context)
generative model 估计的是联合概率分布(joint probability distribution)p()
常见的Generative Model主要有:
– Gaussians, Naive Bayes, Mixtures of multinomials
– Mixtures of Gaussians, Mixtures of experts, HMMs
– Sigmoidal belief networks, Bayesian networks
– Markov random fields
常见的Discriminative Model主要有:
– logistic regression
– SVMs
– traditional neural networks
– Nearest neighbor
Successes of Generative Methods:
NLP
– Traditional rule-based or Boolean logic systems
Dialog and Lexis-Nexis) are giving way to statistical
approaches (Markov models and stochastic context
grammars)
Medical Diagnosis
– QMR knowledge base, initially a heuristic expert
systems for reasoning about diseases and symptoms
been augmented with decision theoretic formulation
Genomics and Bioinformatics
– Sequences represented as generative HMMs
主要应用Discriminative Model:
Image and document classification
Biosequence analysis
Time series prediction
Discriminative Model缺点:
Lack elegance of generative
– Priors, structure, uncertainty
Alternative notions of penalty functions,
regularization, kernel functions
Feel like black-boxes
– Relationships between variables are not explicit
and visualizable
转自:http://cid-2d7821b3af3c6073.spaces.live.com/blog/cns!2D7821B3AF3C6073!160.entry?fl=a