Introduction to Probability (5) Continus random variable

时间:2022-12-18 18:44:08

CONTINUOUS RANDOM VARIABLES AND PDFS 

连续的随机变量,顾名思义。就是随机变量的取值范围是连续的值,比如汽车的速度。气温。假设我们要利用这些參数来建模。那么就须要引入连续随机变量。

假设随机变量X是连续的,那么它的概率分布函数能够用一个连续的非负函数来表示,这个非负函数称作连续随机变量的概率密度函数(probability density function)。并且满足:

Introduction to Probability (5) Continus random variable

假设B是一个连续的区间,那么:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="70" alt="">

要注意的是不论什么一个点的概率是等于零的,由于:

Introduction to Probability (5) Continus random variable

所以对与表示概率时的大于等于。小于等于能够等同于大于和小于:

Introduction to Probability (5) Continus random variable

概率密度函数除了非零这个条件外,另一个条件。依据概率三公理之中的一个的normalization,连续随机变量的总概率等于1:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="65" alt="">

为了直观的理解连续随机变量的概率是什么,例如以下图,连续随机变量在某个区间发生的概率等于该变量概率密度函数在该区间下的面积,如图阴影部分:

Introduction to Probability (5) Continus random variable

所以。对于连续随机变量在区间δ发生的概率为:

Introduction to Probability (5) Continus random variable

直观的表演示样例如以下:

Introduction to Probability (5) Continus random variable

Expectation

连续随机变量X的期望值公式例如以下。就是将离散随机变量中的求和改为了积分:

Introduction to Probability (5) Continus random variable

 

对于随即变量x的函数。其期望值例如以下:

Introduction to Probability (5) Continus random variable

方差例如以下:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="40" alt="">

and:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="40" alt="">

同理离散随机变量,连续随机变量也符合线性原则:

Introduction to Probability (5) Continus random variable

CUMULATIVE
DISTRIBUTION FUNCTIONS 

随机变量的累计概率是指,P(X ≤ x)的概率。表演示样例如以下:

Introduction to Probability (5) Continus random variable

连续随机变量有下面性质:

-单调非递减性:

Introduction to Probability (5) Continus random variable

-FX(x)趋近于0当x趋近于负无穷,FX(x)趋近于1当x趋近于正无穷。

-假设x是离散随机变量。那么FX(x)呈阶梯状上升。假设x是连续随机变量,那么FX(x)呈连续变化上升状。下图分别为离散和连续随机变量的CDF。

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="600" height="160" alt="">

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="600" height="160" alt="">

-假设x是离散随机变量,那么它的PMF能够通过CDF相减得到,CDF能够通过对PMF相加得到:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="80" alt="">

-假设x是连续随机变量,那么它的CDF能够通过对PDF做定积分得到,PDF能够通过对CDF微分得到。

Introduction to Probability (5) Continus random variable

NORMAL RANDOM VARIABLES

正态分布的PDF表演示样例如以下:

Introduction to Probability (5) Continus random variable

μ 是随机变量X的期望,即均值。σ 是随机变量X的标准差。所以方差为σ2

正态分布也满足概率和为一的定理:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="70" alt="">

其PDF和CDF例如以下图所看到的(均值为1,方差为1的正态分布):

Introduction to Probability (5) Continus random variable

当然,正态分布也满足连续随机变量的一般性质:

Introduction to Probability (5) Continus random variable

The Standard Normal Random Variable

标准正态分布是指均值为0,标准差为1的正态分布。它的CDF能够表示为,它的经常使用值被做成了表以供查找:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="400" height="70" alt="">

假设Y等于:Introduction to Probability (5) Continus random variable,那么我们能够将不熟悉的Y转变成X再做计算。公式例如以下:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="34" alt="">

CONDITIONING ON AN EVENT

连续随机变量X与事件A的条件概率表演示样例如以下:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="60" alt="">

类似离散随机变量的条件概率公式,连续随机变量的条件概率例如以下:

Introduction to Probability (5) Continus random variable

连续随机变量X的期望:

Introduction to Probability (5) Continus random variable

对于X的函数g(x)的期望:

Introduction to Probability (5) Continus random variable

相对于离散函数的total probability,连续随机变量也有:

Introduction to Probability (5) Continus random variable

MULTIPLE CONTINUOUS RANDOM VARIABLES

两个连续随机变量的联合分布表演示样例如以下:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="60" alt="">

相同要注意的是f(x,y)是非负的函数。对于一定区间的x,y的概率表演示样例如以下:

Introduction to Probability (5) Continus random variable

像一个随机变量的一样,两个随机变量的PDF满足:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="60" alt="">

为了直观的了解两个随机变量的概念,令:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="400" height="50" alt="">

假设δ无限小,那么双随机变量的概率就相当于是函数f(x,y)在δ2 覆盖下的体积。

连续随机变量的边际概率等于。与离散随机变量的求和相应的是积分:

Introduction to Probability (5) Continus random variableIntroduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="70" alt="">

Expectation

两个随机变量的期望等于:

Introduction to Probability (5) Continus random variable

且有:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="40" alt="">

Conditioning One Random Variable on Another

X,Y是连续随机变量。其联合分布为:fX,Y,  X相对于Y的条件概率为:

Introduction to Probability (5) Continus random variable

条件概率也满足normalization的公式:

Introduction to Probability (5) Continus random variable

期望和条件概率的期望例如以下:

Introduction to Probability (5) Continus random variable

Inference and the Continuous Bayes’ Rule

对于连续的随机变量,也存在贝叶斯准则:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="50" alt="">

对于X是离散随机变量,Y是连续随机变量,贝叶斯准则例如以下:

Introduction to Probability (5) Continus random variable

依据全概率准则,能够得到f(y):

Introduction to Probability (5) Continus random variable

Independence

连续型随机变量和离散型随机变量的独立类似:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="30" alt="">

x与y独立,说明x的发生与否不给y的发生与否提供不论什么信息。反之亦然,那么:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="20" alt="">

Introduction to Probability (5) Continus random variable

假设x,y相互独立,那么他们的乘积的期望等于他们期望的乘积:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="300" height="45" alt="">

另外他们的方差也呈线性:

Introduction to Probability (5) Continus random variable

Joint CDFs

连续随机变量的联合CDF表示为:

Introduction to Probability (5) Continus random variable

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="400" height="50" alt="">

反之。通过二次偏微分能够求得其PDF:

Introduction to Probability (5) Continus random variable

More than Two Random Variables

对于大于两个连续型随机变量的概率公式能够依次类推:

Introduction to Probability (5) Continus random variable

Introduction to Probability (5) Continus random variable

Introduction to Probability (5) Continus random variable

Introduction to Probability (5) Continus random variable

Introduction to Probability (5) Continus random variable

DERIVED DISTRIBUTIONS

对于要求一个连续随机变量的PDF这类问题,我们时常通过绕弯路的方法先求其CDF,再通过对CDF微分求得其PDF。

Introduction to Probability (5) Continus random variable

对于连续随机变量X的线性函数,有:

Introduction to Probability (5) Continus random variable

对于单调函数:

Introduction to Probability (5) Continus random variable

直观的感受是f(X)乘以dh(y)等于P(X),而f(y)乘以dy也等于P(X).例如以下图:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="400" height="350" alt="">

最后回想一下这一章典型的连续型随机变量:

Introduction to Probability (5) Continus random variable

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvZnJvZ19pbl9hX3dlbGw=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" width="600" height="560" alt="">

版权声明:本文博主原创文章。博客,未经同意不得转载。

Introduction to Probability (5) Continus random variable的更多相关文章

  1. Codeforces Round #348 (VK Cup 2016 Round 2, Div. 1 Edition) C. Little Artem and Random Variable 数学

    C. Little Artem and Random Variable 题目连接: http://www.codeforces.com/contest/668/problem/C Descriptio ...

  2. 【概率论】4-1:随机变量的期望(The Expectation of a Random Variable Part II)

    title: [概率论]4-1:随机变量的期望(The Expectation of a Random Variable Part II) categories: - Mathematic - Pro ...

  3. 【概率论】3-8:随机变量函数(Functions of a Random Variable)

    title: [概率论]3-8:随机变量函数(Functions of a Random Variable) categories: Mathematic Probability keywords: ...

  4. 【概率论】4-1:随机变量的期望(The Expectation of a Random Variable Part I)

    title: [概率论]4-1:随机变量的期望(The Expectation of a Random Variable Part I) categories: - Mathematic - Prob ...

  5. Jmeter入门16 数据构造之随机数Random Variable & __Random函数

     接口测试有时参数使用随机数构造.jmeter添加随机数两种方式 1  添加配置 > Random Variable  2  __Random函数   ${__Random(1000,9999) ...

  6. Simple Random Sampling|representative sample|probability sampling|simple random sampling with replacement| simple random sampling without replacement|Random-Number Tables

    1.2 Simple Random Sampling Census, :全部信息 Sampling: 抽样方式: representative sample:有偏向,研究者选择自己觉得有代表性的sam ...

  7. Fuzzy Probability Theory---(3)Discrete Random Variables

    We start with the fuzzy binomial. Then we discuss the fuzzy Poisson probability mass function. Fuzzy ...

  8. Random variable

    \(\underline{cdf:}\)cumulative distribution function \(F(x)=P(X \leq x)\) \(\underline{pmf:}\)probab ...

  9. Introduction to Probability (三) Independence

    两个事件独立性的定义是:事件A的发生对事件B的发生毫无影响,即从A的发生与否.我们不能猜測出B是否发生. 从概率等式的表示来看就是B在A发生的情况下发生的概率等于B发生的概率本身. 进而引出了A与B同 ...

随机推荐

  1. [蟒蛇菜谱] Python封装shell命令

    # -*- coding: utf-8 -*- import os import subprocess import signal import pwd import sys class MockLo ...

  2. ehcache2拾遗之cache持久化

    问题描述 应用在使用过程中会需要重启等,但是如果ehcache随着应用一起重启,那么刚重启的时候就会出现大量的miss,需要一定的访问量来重建缓存,如果缓存能够持久化,重启之后可以复用将会有助于缓解重 ...

  3. 图片代替radio

    <html> <head> <meta http-equiv="Content-Type" content="text/html; char ...

  4. WPF 蒙层罩,正在加载

    参考园子里的一篇文章,比较好用.可以直接用,可以自己改. 动画效果: 容器的触发器,旋转容器: 属性配置:使用依赖属性,并且在xaml中写绑定.

  5. Linux 命令整理 —— 用户管理

    Linux用户管理以读.写.执行动作为权限,以用户组为单位,限制用户行为.对于文件的的操作,可以限制读.写.执行中的哪一种,也可以限制文件所有者.组用户.组外用户相应的权限. 所以,要建立用户,最好先 ...

  6. OJ提交题目中的语言选项里G&plus;&plus;与C&plus;&plus;的区别(转载)

    原文链接:http://blog.polossk.com/201405/c-plus-plus-g-plus-plus G++? 首先更正一个概念,C++是一门计算机编程语言,G++不是语言,是一款编 ...

  7. Oracle,MySQL,sqlserver三大数据库如何获取系统当前时间

    Oracle中如何获取系统当前时间:用SYSDATE() MySQL中获取系统当前时间主要有以下几点: (1)now()函数以('YYYY-MM-dd HH:mm:SS')返回当前的日期时间,可以直接 ...

  8. process credentials&lpar;一&rpar;

    一.介绍 当linux系统中的一个进程运行起来的时候,总是要访问系统的资源,访问文件或者向其他的进程发送信号.系统是否允许其进行这些操作?系统是根据什么来判断该进程的权限?这些问题是和进程信任状(pr ...

  9. sublime text--你所不知道的12个秘密

    转自:http://blog.csdn.net/laokdidiao/article/details/51429404 package control安装: 第一种方法是在控制台中复制粘贴代码按回车, ...

  10. os层删除与数据库层drop的区别于对策。

    os rm删除的,可以通过os层面恢复(句柄.inode.如果inode没有了可以根据block去重组出来数据文件),也可以通过rman备份还原方式恢复.  db 层面drop删除的 只能不完全恢复( ...