QM5_Didstribution
Basic Concepts
Probability distribution
Discrete distribution (离散分布)
- The distribution of the discrete random variable.
- Discrete random variable
- takes on a finite and countable number of possible values.
Continuous distribution (连续分布)
- The distribution of the continuous random variable.
- Continuous random variable:
- takes on an infinite and uncountable number of possible values.
Probability function
Probability function (概率函数)
- For discrete random variable taking on a specific value
- p(x) = P(X=x)
- X: 随机变量; x: a specific value
Probability density function (概率密度函数)
- For continuous random variables within a range
- P(x1<X<x2)
- The probability of taking on an specific value is always zero, P(X=x)=0
- 因为连续随机变量有无数个数, 即分母为无穷大, 所以取到一个具体的值的概率是0. 这并不代表不可能取到个这值,只是相对于取值范围, 取到该值的可能性太小. 所以对连续随机变量, 研究具体的值没有什么价值, 研究的是range.
Cumulative probability function (累积概率函数)
- F(x) = P(X<=x)
插入图片, 用面积代表概率.
Dsicrete uniform distribution (离散均匀分布)
- Definition
- has a finite number of possible outcomes, all of which are equally likely.
- Example: X = {1,2,3,4,5}
- P(1) = P(2) = P(3) = P(4) = P(5) = 20%
- P(3) = P(1) + P(2) + P(3) = 60%
- P(2<=X<=4) = P(2) + P(3) + P(4) = 60%
Discrete Distribution
Binomial distribution ** (二项式分布)
Bernoulli random variable (trial,伯努利实验)
- Random variables with only two outcomes, one represents success(denoted as 1); the other represents failure(denoted as 0). P(X=1) = p, P(X=0) = 1-P.
- Binomial random variable
- The number of successes in a Bernoulli trials. (做n次Bernoulli trials就得到二项式分布)
- The probability of x successes in n trails.
插入老师板书.插入公式.
- Expected value and variance
插入图片
- 均值 => 期望值 => 算加权平均
- 计算器算排列组合: 10个中挑出6个, 10 => 2nd => + ->6 -> =.
Continuous Distribution
Continuous uniform distribution (连续均匀分布)
- Definition
- probability of continuous uniform random variable which distribute evenly over an interval.
- Properties
- P(X=x)=0
- P(x1<=X<=x2) = (x2-x1)/(b-a)
Normal distribution *** (正态分布)
Properties
- completely described by mean and variable.
- 只由两个参数决定 , 均值和方差.
插入公式
skewness = 0, kurtosis = 3
- Linear combination of normally distributed random variables is also normally distributed.
- 比如x1~n是线性的, x2~n也是线性的, 则3x1+2x2~n也是线性的.
- Probability descrase further from the mean, but the tails go on forever.
- completely described by mean and variable.
考点
- 性质
- 置信区间
- 标准化
Concepts
- Confidence interval 置信区间
- 落在区间内的概率, 就是切比雪夫.
- Confidence level 置信水平
- 置信水平 = 置信度
- Confidence degree 置信度
- Confidence interval 置信区间
- Properties
插入图
k : 依赖因子(关键值), Reliability factor /Critical value.
Standard normal distribution 标准正态分布
- also named z-distribution
- X~N (0,1), 正态分布均值是0, 方差是1.
- Standardization
- >>>插入公式
- z值含义
- 正态分布与标准正态分布位置相对应.
- 离标准正态分布均值的距离,即z个标准差.
- 算出z值后,查表(z-table), 查出累计概率.
- 例题
Shortfall risk **(缺口风险)
- Definition
- the risk that portfolio value or return will fall below the imnimum acceptable level(RL)
- Properties
- The lower, the better
Safety-first ratio **(第一安全比率)
- Definition
- the distance from the mean return to the shortfall elvel in units of standard deviation.
- Calculation
- >>>插入公式
- Properties
- The higher, the better
- Minimizing shortfall risk - Maximizing safety-first ratio
Lognormal distribution **
- Properties
插入图片
Student;s t-distribution ***
- Properties
- Defined by single parameter: degree of freedom(df), 由唯一一个参数决定.
- df = n-1, where: n is the sample size.
- Symmetrical, skewness = 0,
- Fatter tails than a normal distribution (低峰肥尾)
- As df increase t-distribution is approaching to standard normal distribution.
- Given a degree of confidence, t-distribution has a wider confidence interval than z-distribution.
- Defined by single parameter: degree of freedom(df), 由唯一一个参数决定.
- Shape
插图
Simulation (模拟)
插入图片
QM5_Didstribution的更多相关文章
随机推荐
- SharePoint 2007 文档库中的文档添加评论功能
背景:接到一个项目,要求文档管理,当然文档库就可以了,但是要求文档需要大家去读,读完以后还可以发表评论,这Moss貌似就有点困难了.和同事一起合计,想来想去也没有太好的办法,后来想到传统开发,两个表的 ...
- CRF资料
与最大熵模型相似,条件随机场(Conditional random fields,CRFs)是一种机器学习模型,在自然语言处理的许多领域(如词性标注.中文分词.命名实体识别等)都有比较好的应用效果.条 ...
- 初探linux子系统集之led子系统(一)
就像学编程第一个范例helloworld一样,学嵌入式,单片机.fpga之类的第一个范例就是点亮一盏灯.对于庞大的linux系统,当然可以编写一个字符设备驱动来实现我们需要的led灯,也可以直接利用g ...
- 十大ios开发者喜爱的开源库
十大ios开发者喜爱的开源库 (转自博客园) 2014-08-17 14:07:58| 分类: objective-c | 标签:ios 开源库 |举报|字号 订阅 下载LOFTER我的照片书 ...
- obj-c编程16:键值编码(KVC)
我们可以借助obj-c中的键值编码(以后简称KVC,Key-Value Coding)来存取类的属性,通过指定所要访问的属性名字符串标示符,可以使用存取器方法来获取或设置类的属性.下面的例子,借助于K ...
- Asp.NetCore+Microsoft.AspNetCore.SignalR前后端分离
1.新建WebApi 2.安装Microsoft.AspNetCore.SignalR 3.新建一个集线器和消息类 using Microsoft.AspNetCore.SignalR; using ...
- ValueObject的理解
思考ValueObject应该更多从内存的角度思考,而非DB持久化的角度. 例如: public class A { public int Id { get; set; } public Addres ...
- 初识Java——一维数组的创建及使用
数组作为对象是允许使用new关键字进行内存分配的,在使用数组前,必须首先定义数组的变量所属的类型.一维数组的创建有两种方法: 1,先声明,再用new运算符进行内存分配 数组元素类型+数组名字[] 数组 ...
- Python中的函数与变量
本节内容 函数的定义方法 函数功能 函数的返回值 函数的形参与实参 全局变量与局部变量 递归 函数的作用域 匿名函数lambda 函数式编程 常用内置函数 其他内置函数 函数 函数的定义方法 函数就相 ...
- Gradle初探
(一):创建一个Gradle项目 1. 环境准备 1.1. 先上Gradle官网下载最新版的程序,地址:https://gradle.org/gradle-download/. 1.2. 配置环境变量 ...