Mean, Variance, and Covariance
βνλ₯ κ³Ό ν΅κ³(MATH230)β μμ μμ λ°°μ΄ κ²κ³Ό 곡λΆν κ²μ μ 리ν ν¬μ€νΈμ λλ€. μ 체 ν¬μ€νΈλ Probability and Statisticsμμ νμΈνμ€ μ μμ΅λλ€ π²
Mean
Definition.
The <expectation> or <mean> of a RV $X$ is defined as
\[\mu := E[x] := \begin{cases} \displaystyle \sum_x x f(x) && X \; \text{is a discrete with pmf} f(x) \; \\ \displaystyle \int^{\infty}_{\infty} x f(x) dx && X \; \text{is a continuous with pdf} \; f(x) \end{cases}\]λ§μ½ RV $X$μ ν¨μ $g(x)$λ₯Ό μ·¨νλ€λ©΄, <Expectation>μ μλμ κ°μ΄ ꡬν μ μλ€.
Theorem.
Let $X$ be a random variable with probability distribution $f(x)$. The expected value of the random variable $g(X)$ is
\[\mu_{g(X)} = E\left[g(X)\right] = \sum_x g(x) f(x) \quad \text{if } X \text{ is discrete RV}\]and
\[\mu_{g(X)} = E\left[g(X)\right] = \int^{\infty}_{\infty} g(x) f(x) \quad \text{if } X \text{ is continuous RV}\]($g(x)$λ₯Ό μ·¨νλ μ¬μ ν $x$μ μ μμμ μ μ§λλ―λ‘, μμ κ°μ΄ $g(x) f(x)$λ₯Ό μ¬μ©νλ κ²μ νλΉνλ€.)
ps) μμ μκ°μ κ΅μλκ»μ μ΄μ° RVμ λν μ¦λͺ μ μ½κ² ν μ μμ§λ§, μ°μ RVμ λν μ¦λͺ μ μ’ κΉλ€λ‘λ€κ³ νμ ¨λ€.
μ΄λ²μλ joint distributionsμ λν <Expectation>μ μ΄ν΄λ³΄μ.
Definition.
Let $X$ and $Y$ be RVs with joint probability distribution $f(x, y)$. The expected value of the RV $g(X, Y)$ is
\[\mu_{g(X, Y)} = E\left[g(X, Y)\right] = \sum_x \sum_y g(x, y) f(x, y) \quad \text{if } X \text{ and } Y \text{ is discrete RV}\] \[\mu_{g(X, Y)} = E\left[g(X, Y)\right] = \int^{\infty}_{-\infty} \int^{\infty}_{-\infty} g(x, y) f(x, y) \; dx dy \quad \text{if } X \text{ and } Y \text{ is continuous RV}\]Conditional Distributionμ λν΄μλ <Expectation>μ μκ°ν΄λ³Ό μ μλ€.
Definition.
Linearity of Expectation
<Expectation>μ <Linearity>λΌλ μμ£Ό μ’μ μ±μ§μ κ°μ§λ€.
Theorem.
Let $a, b \in \mathbb{R}$, then $E\left[aX + b\right] = aE[X] + b$.
μμ μ λ¦¬κ° λ§ν΄μ£Όλ κ²μ <Expectation>μ΄ Linear Operatorμμ λ§ν΄μ€λ€!! π€©
μ’λ νμ₯ν΄μ κΈ°μ ν΄λ³΄λ©΄,
Theorem.
Theorem.
Expectation with Independence
λ§μ½ λ RV $X$, $Y$κ° μλ‘ <λ 립>μ΄λΌλ©΄, λ RVμ κ³±μ λν <Expectation>μ μ½κ² ꡬν μ μλ€.
Theorem.
If $X$ and $Y$ are independent, then
\[E[XY] = E[X]E[Y]\]Variance and Covariance
λ RV $X$, $Y$κ° λμΌν νκ· μ κ°μ§λλΌλ; $E[X] = \mu = E[Y]$ RVμ κ°λ³ κ°λ€μ΄ νκ· $\mu$λ‘λΆν° λ¨μ΄μ Έ μλ μ λλ λ€λ₯Ό μ μλ€. <λΆμ° Variance>λ μ΄λ° νκ· μΌλ‘λΆν°μ νΌμ§ μ λλ₯Ό μΈ‘μ νλ μ§νλ‘ μλμ κ°μ΄ μ μνλ€.
Definition.
The <variance> of a RV $X$ is defined as
\[\text{Var}(X) = E[(X-\mu)^2]\]and $\sigma = \sqrt{\text{Var}(X)}$ is called the <standard deviation> of $X$.
μλμ 곡μμ μ¬μ©νλ©΄, $\text{Var}(X)$λ₯Ό μ’λ μ½κ² ꡬν μ μλ€.
Theorem.
βλΆμ° = μ ν - νμ β, κ³ λ±νκ΅ λ λ°°μ΄ κ³΅μμ΄λ€!
<Expectation>μ LinearityλΌλ μ’μ μ±μ§μ κ°μ§κ³ μμλ€. <λΆμ° Variance>μμλ μ΄λ»κ² λλμ§ μ΄ν΄λ³΄μ.
Theorem.
For any $a, b \in \mathbb{R}$,
\[\text{Var}(aX + b) = a^2 \text{Var}(X)\]Covariance
<곡λΆμ° Covariance>λ λ RV μ¬μ΄μ μ΄λ€ <κ΄κ³ relation>μ΄ μλμ§λ₯Ό μ‘°μ¬νλ μ§νλ€. <곡λΆμ°>μ μλμ κ°μ΄ μ μλλ€.
Definition.
The <covariane> of $X$ and $Y$ is defined as
\[\begin{aligned} \sigma_{XY} := \text{Cov}(X, Y) &= E \left[ (X - \mu_X) (Y - \mu_Y) \right] \\ &= E(XY) - E(X)E(Y) \end{aligned}\]- $\text{Cov}(X, X) = \text{Var}(X)$
- $\text{Cov}(aX + b, Y) = a \cdot \text{Cov}(X, Y)$
- $\text{Cov}(X, c) = 0$
μμμ μ΄ν΄λ΄€μ λ, λ RV $X$, $Y$κ° λ 립μ΄λΌλ©΄, $E(XY) = E(X)E(Y)$κ° λμλ€. λ°λΌμ λ RVκ° λ λ¦½μΌ λλ $\text{Cov}(X, Y) = 0$μ΄ λλ€! κ·Έλ¬λ μ£Όμν μ μ λͺ μ μ μ(ζ)μΈ $\text{Cov}(X, Y) = 0$μΌ λ, λ RVκ° νμ λ 립μμ 보μ₯νμ§λ μλλ€!
<Covariance>μ λ RVμ Linear Combinationμ λν λΆμ°μ ꡬν λλ μ¬μ©νλ€.
Let $a, b, c \in \mathbb{R}$, then
\[\text{Var}(aX + bY + c) = a^2 \text{Var}(X) + b^2 \text{Var}(Y) + 2 \text{Cov}(X, Y)\]μ¦λͺ μ $\text{Var}(aX + bY + c)$μ μλ―Έλ₯Ό κ·Έλλ‘ μ κ°νλ©΄ μ½κ² μ λν μ μλ€.
\[\text{Var}(aX + bY + c) = E\left[ \left( (X+Y) - (\mu_X + \mu_Y) \right)^2 \right]\]Correlation
<곡λΆμ°>μ μ’λ 보기 μ½κ² Normalize ν κ²μ΄ <Correlation>μ΄λ€.
Definition.
The <correlation> of $X$ and $Y$ is defined as
\[\rho_{XY} := \text{Corr}(X, Y) = \frac{\text{Cov}(X, Y)}{\sqrt{\text{Var}(X)} \sqrt{\text{Var}(Y)}}\]- if $\rho_{XY} > 0$, $X$ and $Y$ are positively correlated.
- if $\rho_{XY} < 0$, $X$ and $Y$ are negatively correlated.
- if $\rho_{XY} = 0$, $X$ and $Y$ are uncorrelated.
λ§μ½ λ RVκ° μλ²½ν μ νμ±μ 보μΈλ€λ©΄, $\rho_{XY}$κ° μλμ κ°λ€.
- if $Y = aX + b$ for $a > 0$, then $\text{Corr}(X, Y) = 1$
- if $Y = aX + b$ for $a < 0$, then $\text{Corr}(X, Y) = -1$
μμ λͺ μ λ κ·Έ μλ μ±λ¦½νλ€. μ¦λͺ μ μλμ Exerciseμμ μ§ννκ² λ€.
<Correlation>μ $[-1, 1]$μ κ°μ κ°λλ€. μ΄λ <μ½μ-μλ°λ₯΄νΈ λΆλ±μ>μ ν΅ν΄ μ λν μ μλ€!
Cauchy-Schwarrtz inequality :
\[\left( \sum a_i b_i \right)^2 \le \sum a_i^2 \sum b_i^2\]Correlation μμ μλ―Έμ λ§‘κ² νμ΄μ°λ©΄ μλμ κ°λ€.
\[\begin{aligned} \text{Corr}(X, Y) &= \frac{\text{Cov}(X, Y)}{\sqrt{\text{Var}(X)} \sqrt{\text{Var}(Y)}} = \frac{E[(X-\mu_X)(Y - \mu_Y)]}{\sqrt{E[(X-\mu_X)^2]} \sqrt{E[(Y-\mu_Y)^2]}} \\ &= \frac{\sum (X-\mu_X)(Y - \mu_Y)}{\sqrt{\sum (X-\mu_X)^2} \sqrt{\sum (Y-\mu_Y)^2}} \end{aligned}\]μ΄μ μμ μμ μ κ³±ν΄μ μ΄ν΄λ³΄λ©΄
\[(\rho_{XY})^2 = \left( \frac{\sum (X-\mu_X)(Y - \mu_Y)}{\sqrt{\sum (X-\mu_X)^2} \sqrt{\sum (Y-\mu_Y)^2}} \right)^2 = \frac{\left( \sum (X-\mu_X)(Y - \mu_Y) \right)^2 }{\sum (X-\mu_X)^2 \sum (Y-\mu_Y)^2}\]<μ½μ-μλ°λ₯΄μΈ λΆλ±μ>μμ μ°λ³μ μ’λ³μΌλ‘ μ΄λνλ©΄, μλμ κ°μ λΆλ±μμ΄ μ±λ¦½νλ€.
\[\frac{\left( \sum a_i b_i \right)^2}{\sum a_i^2 \sum b_i^2} \le 1\]μ΄λ₯Ό <Correlation>μ μ κ³±μμ μ μ©νλ©΄ μλμ κ°λ€.
\[(\rho_{XY})^2 = \frac{\left( \sum (X-\mu_X)(Y - \mu_Y) \right)^2 }{\sum (X-\mu_X)^2 \sum (Y-\mu_Y)^2} \le 1\]λ°λΌμ $(\rho_{XY})^2 \le 1$μ΄λ―λ‘
\[-1 \le \rho_{XY} \le 1\]$\blacksquare$
μΆκ°λ‘ <Correlation>μ βνμ€νβν RVμ 곡λΆμ°μΌλ‘λ ν΄μν μ μλ€.
$Z = \dfrac{X-\mu_X}{\sigma_X}$, $W = \dfrac{Y-\mu_Y}{\sigma_Y}$λΌκ³ νμ€ννλ€λ©΄, μ΄ λμ 곡λΆμ°μ $X$, $Y$μ λν Correlationκ³Ό κ°λ€.
\[\text{Var}(Z, W) = \text{Corr}(X, Y)\]λ± λ³΄λ©΄ μ¦λͺ ν μ μμ κ² κ°μμ λ°λ‘ μ λλ νμ§ μκ² λ€.
Q1. $\text{Var}(X) = 0$λ 무μμ μλ―Ένλκ°?
A1.
Q2. $\text{Cov}(X, Y) = 0$μ΄μ§λ§, λ RVκ° λ λ¦½μ΄ μλ μλ₯Ό μ μνλΌ.
Q3. Prove that $-1 \le \text{Corr}(X, Y) \le 1$.
Q4. Prove that if $\text{Corr}(X, Y) = 1$, then there exist $a>0$ and $b\in\mathbb{R}$ s.t. $Y = aX + b$.
νΌμ³λ³΄κΈ°
A1. $p(x)$κ° delta-functionμμ μλ―Ένλ€.
A2. $Y=X^2$μΌλ‘ μ€μ νλ©΄ μ½κ² λ³΄μΌ μ μλ€. λ 립μμ 보μ΄κΈ° μν΄ $p(x, y)$λ₯Ό ꡬν΄μΌ ν μλ μλλ°, μ΄κ² μμ μ μ ν μ μ€μ ν΄μ£Όλ©΄ μ½κ² reasonableνκ² λμμΈ ν μ μμ κ²μ΄λ€.
A3. & A4. Q3λ μ΄λ―Έ μμμ μ¦λͺ μ νλ€. κ·Έλ¬λ λ€λ₯Έ λ°©μμΌλ‘λ μ¦λͺ ν μ μλ€! π μ΄κ³³μ [2, 3]pλ₯Ό μ°Έκ³ νλΌ.
μ΄μ΄μ§λ λ΄μ©μμλ <νκ· >κ³Ό <λΆμ°>μ λν μ½κ°μ μΆκ°μ μΈ λ΄μ©μ μ΄ν΄λ³Έλ€.
κ·Έλ¦¬κ³ Discrete RVμμμ κΈ°λ³Έμ μΈ Probability Distributionμ μ΄ν΄λ³Έλ€.
- Bernoulli Distribution
- Binomial Distributions
- Multinomial Distribution
- Hypergeometric Distributions
- etcβ¦