Home » , , » Generative Model and Discrimitive Model

# Generative Model and Discrimitive Model

Not very easy to distinguish between them, here I just give some useful link to this two kinds of model.

For a data sample: x and it class lable:,y,

Discrimitive model: p(y|x), 给定x,算y。
Generative model:和 discrimitive model 的区别：
Generative model focus

Generative model 实际上带的information 要比discrimitive model rich,

decision boundary。这里说的generative model 和 discrimitive
model，在行业里，这个说法是通用的。

generative model。因为需要用到P(x), 如果只是label 的话，p(y)很简单，但为什么不能直接用Generative
model 呢？

Discrimitive model: 相当于在图像上scan 一下，detection, 用一个path, 在不同的scale

Discrimitive 比较easy to learn， 给出正负例，给出lable, focus on discrimitive
model marginal distribution。 某种意义上，比generativemodel 要简单，但power 是
limited, 可以告诉你的时1还是2，但没有办法把整个场景描述出来。
P(x|y) P(y)

computing 要比 discrimitive model 复杂得多

Discriminative Model是判别模型，又可以称为条件模型，或条件概率模型。
Generative Model是生成模型，又叫产生式模型。

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()

– Gaussians, Naive Bayes, Mixtures of multinomials
– Mixtures of Gaussians, Mixtures of experts, HMMs
– Sigmoidal belief networks, Bayesian networks
– Markov random fields

– logistic regression
– SVMs
– 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

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
Bridging Generative and Discriminative：
Can performance of SVMs be combined
elegantly with flexible Bayesian statistics?
Maximum Entropy Discrimination marries
both methods
– Solve over a distribution of parameters (a
distribution over solutions)