术语
| 术语 | 中文含义 | 备注 |
| Bayesian inference | ||
| density estimation | ||
| ill-posed | 不适定的 | # |
| parametric distribution | ||
| conjugate prior | 共轭先验 | * |
| exponential family | Dirichlet, Gaussian… | |
| Dirichlet distribution | ||
| Gaussian distribution | 高斯分布, 正态分布 | |
| nonparametric density estimation | histograms, nearest-neighbours, kernels | |
| histograms | ||
| nearest-neighbours | ||
| kernels |
注：#(不确定含义)，*(重点)
本章用到的假设
Data points are independent and identically distributed(i.i.d.)
数据点都是独立且同分布的。
重点
Density estimation is fundamentally ill-posed, because there are infinitely many probability distributions that could have given rise to the observed finite data set.
密度估计从根本上讲是不适定的，因为有无穷多的概率分布能满足已观测到的有限的数据集。
Conjugate priors lead to posterior distributions having the same functional form as the prior, and that therefore lead to a greatly simplified Bayesian analysis.
共轭先验使得后验分布同先验分布有着相同的函数形式，因此大大简化了贝叶斯分析。
One limitation of the parametric approach: It assumes a specific functional form for the distribution, which may turn out to be inappropriate for a particular application.
参数化方法的一个局限：它假设分布会有一个具体的函数形式，但这个函数形式可能与实际情况不符。
Nonparametric methods in which the form of the distribution typically depends on the size of the data set. Such models still contain parameters, but these control the model complexity rather than the form of the distribution.
非参数化方法中分布的形状通常取决于数据集的大小，此类模型虽然也有参数，但这些参数主要控制模型的复杂度而非分布的形状。
-------------------------------------------
作者：兔纸张 来源：博客园 ( http://www.cnblogs.com/geiliCode )
