PRML阅读记(2)

继续PRML
第二章:
Binary Variables:
  • Bernouli Distribution, binomial distribution
  • conjugate prior --> beta distribution
Multinomial Variables:
  • multinomial distribution
  • conjugate prior --> Dirichlet distribution
The Gaussian Distribution:
  • univariate, multivariate, shape, limit
  • conditional Gaussian, Marginal Gaussian, Bayes' theorem
  • Maximum likelihood, sequential estimation
  • conjugate prior --> unknown mean, unknown variance, and both
  • Mixtures of Gaussians
The exponential Family:
  • Bernoulli distribution --> logistic sigmoid function
  • Multinomilal distribution --> softmax function
  • conjugate priors
Nonparametric Methods:
  • histogram approach: p(i) = n(i) / (N * width of bin)
  • p(x) = K/NV:   fix V and find K --> Kernel Estimator . fix K and find V --> K-nearest-nerghbour estimator
  • Kernel Estimator: estimate new data x according to old datas in V
  • K-nearest-nerghbour estimator: estimate new data x according to neighbors within K

import views:
1. posterior = prior * ML
2. conjugate prior
3. sequential model to deal with large dataset(update data with disgarding the old data)
4. Gaussian Distribution and its variation
5. nonparametic method
6. hyperparameter: to model the distribution of parameter
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值