统计模型和概率模型之间的区别?


29

应用概率是概率的重要分支,包括计算概率。由于统计是使用概率论来构建处理数据的模型,据我所知,我想知道统计模型与概率模型之间的本质区别是什么?概率模型不需要真实数据吗?谢谢。

Answers:


29

概率模型由三联体,其中,Ω是样本空间,˚Fσ代数(事件)和P是一个概率测度˚F(Ω,F,P)ΩFσPF

直观的解释。一种概率模型可以解释为一个已知的 随机变量 。例如,令X为均值0和方差1的正态分布随机变量。在这种情况下,概率测度P与相关联的累积分布函数(CDF)˚F通过XX01PF

F(x)=P(Xx)=P(ωΩ:X(ω)x)=x12πexp(t22)dt.

概括。概率模型的定义取决于概率的数学定义,例如,参见自由概率量子概率

统计模型 是一概率模型,这是一组概率测度/样品上空间分布ΩSΩ

通常选择这组概率分布来对我们拥有数据的某种现象进行建模。

直观的解释。在统计模型中,描述某种现象的参数和分布都是未知的。这样的一个例子是正态分布的与家庭内的平均和方差σ 2[R +,这是,这两个参数是未知的并且您通常要使用的数据集,用于估计参数(即,选择的元件小号)。可以在任何ΩF上选择这组分布,但是,如果我没有记错的话,在实际示例中,仅在同一对上定义的分布Ω FμRσ2R+SΩF(Ω,F) 有合理的考虑。

概括本文提供了统计模型的非常正式的定义,但是作者提到“贝叶斯模型需要先验分布形式的附加组成部分……尽管贝叶斯公式不是本文的主要重点”。因此,统计模型的定义取决于我们使用的模型类型:参数模型或非参数模型。同样在参数设置中,定义取决于如何处理参数(例如,古典与贝叶斯)。

区别是:在一个概率模型你确切地知道的概率的措施,例如一个,其中,μ 0σ 2 0是已知的参数,而在统计模型你考虑集分布。中,例如普通μ σ 2,其中μ σ 2是未知参数。Normal(μ0,σ02)μ0,σ02Normal(μ,σ2)μ,σ2

它们都不要求数据集,但我要说的是通常选择统计模型来建模。


2
@HonglangWang在某种程度上是正确的。主要区别在于,概率模型只是一个(已知)分布,而统计模型是一组概率模型。数据用于从该集合中选择一个模型,或者从模型(根据数据)更好地(在某种意义上)描述现象的较小模型子集。

2
(+1) This is a nice answer, though I have a couple of comments. First, I think this may be selling the probabilist a little bit short. It is not at all uncommon to consider a set of probability spaces in a probabilistic model, and indeed, the possible measures can even be random (constructed on a suitably larger space). Second, a Bayesian (in particular) might find this answer slightly disconcerting in that a Bayesian statistical model can often be viewed as a single probability model on a suitable product space Ω×Θ.
cardinal

1
PΩP(Xx)P(ωΩ:X(ω)x), then Ω are not observable values. F is a σalgebra which is the pre-image of the Borel σalgebra under X, again this are not observable. I am not sure how to explain this in an intuitive level.

2
@gung Ω depends on the application; it is not determined by theory. For instance, Ω could be a set of Brownian motions describing the price of a financial derivative and X could be the value attained at a fixed time t. In another application Ω could be a set of people and X could be the lengths of their forearms. Generally, Ω is a mathematical model of the physical objects of study and X is a numerical property of those objects. F is the set of possible events: those situations to which we want to ascribe probabilities.
whuber

2
@gung F is a sigma algebra: it's a collection of subsets (the "events"). In the financial application, it's a set of price histories; in the forearm measurements application, the events would be sets of people. We can talk about this more if you want in a chat room.
whuber
By using our site, you acknowledge that you have read and understand our Cookie Policy and Privacy Policy.
Licensed under cc by-sa 3.0 with attribution required.