II-Descriptive Statistics-5 Flashcards

1
Q

4种Measurement of scales

A
  • nominal scales
  • ordinal scales (大小)
  • interval scales (加减,零)
  • ratio scales(乘除,零)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

总体

样本

总体参数

样本统计量

A

pupolation

sample

population parameter

sample statistic

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Frequency Distribution

A

频率分布 以图表方式展示统计数据的方法 对所有衡量尺度都适用

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Absolute Frequency

Relative Frequency

Cumulative Frequency

Cumulative relative Frequency

A
  1. 频率分布中每一区间里 观测值的数量
  2. AF/total number of observations
  3. frequencies add up to now
  4. frequencies add up to now / total number of observations
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

modal interval

A

众数区间,有最多观测值的区间

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

3 kinds of graphic presentation of

Frequency Distribution

A
  • histogram frenquency 直方图
  • polygon 折线图
  • cumulative frequency distribution
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

6 means of central tendency

A
  • arithmetic: population and sample
  • median
  • mode
  • weighed
  • geometric
  • harmonic
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

arithmetic means的3个特点

A
  1. 所有观测值都被用到
  2. 唯一一个偏差总和为零的度量
  3. 缺陷:被极大的偏差带跑
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

mode中位数取值方法

A

1.观测值排序

2,奇数个,取中间值,偶数个,取n/2 ; (n+2)/2两个值的平均数

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

harmonic mean

A

描述平均价格

$1000可以买多少8股票,多少9,多少10? 除以3,得出平均可以买多少股票。再用$1000除以此数,得出平均价格

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

geometric mean harmonic mean arithmetic mean 三者关系

Weighted mean呢?

A

harmonic mean小于geometric mean小于arithmetic mean

Weighted mean无法参与比较,因为权重不同结果不同

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

对于基金经理

衡量过去业绩用哪种mean

预测未来业绩用哪种mean

A

过去用geometric

未来用arithmetic

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

四种quantile

A

分位数

  • quartile
  • quintile
  • decile
  • percentile
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Quantile:

  1. equation
  2. quantile = 7.2含义
A
  • Ly = (n + 1) * ?%
  • obervations中第7第8数字中间,距离第7, 20% (linear interpolation线性插值)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

计算quantile时应注意什么

Third Quartile代表什么

A
  1. 所有观测值升序排列
  2. third quartile即从上往下75%
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

measures of

absolute dispersion

及算法

A
  • range:max-min
  • mean absolute deviation (MAD):每一观测值与算术平均值之差的绝对值相加求平均
  • variance:观测值与平均值的差值平方,总体除N,样本除N-1
  • standard deviation:方差开根号
17
Q

MAD与standard deviation的关系

A

standard diviation

总是大于等于

MAD

18
Q

semivariance/semideviation

target semiviarance / target semideviation

算法

A
  1. 求平均值
  2. 找出低于或等于平均值(或目标值)的观测值
  3. 求方差,除以观测值数量N-1
19
Q

字母辨析

S

σ

Xbar

μ

A
  • 样本标准差
  • 总体标准差
  • 样本均值
  • 总体均值
20
Q

Chebyshev’s inequality

definition

A
  1. 一组方差为有限值的样本,
  2. K (>1)个标准差范围内观测值所占的百分比为

1 - 1 / k2

21
Q

2 measurements of relative dispersion

A
  1. coefficient of variation
  2. Sharpe Ratio
22
Q

coefficient of variation CV

equation, and

what it means

A

the amount of risk per unit of return

23
Q

Sharpe Ratio:

equation, and

what it means

A

mean excess return per unit risk

24
Q

使用sharpe ratio时

应注意的问题

A
  1. SR有可能是负数,负数时风险(分母)越大,SR值反而越大,因此不可理解为SR越大,投资表现一定更好
  2. SR基于标准差计算,标准差适于描述对称分布,对于skewed分布描述不准确
25
Q

skewness: equation

准确版和简易版

A

标准版:

系数为 1 / (n-1)(n-2)

26
Q

mode median mean三者关系

  1. positively skewed
  2. negatively skewed
A
  1. mode < median < mean
  2. mode > median > mean
27
Q

kurtosis

几种描述词汇

A

峰度

keptokurtic尖峰

platykurtic平峰

mesokurtic正常分布

28
Q

excess kurtosis

equation

简易版

A

以下公式 - 3 (正态分布峰度)

29
Q

geometric mean

arithmetic mean

量化关系

A

geo = arith - (variance of return /2)