$\gdef \vec#1{\boldsymbol{#1}} \\ \gdef \rank {\mathrm{rank}} \\ \gdef \det {\mathrm{det}} \\ \gdef \Bern {\mathrm{Bern}} \\ \gdef \Bin {\mathrm{Bin}} \\ \gdef \Mn {\mathrm{Mn}} \\ \gdef \Cov {\mathrm{Cov}} \\ \gdef \Po {\mathrm{Po}} \\ \gdef \HG {\mathrm{HG}} \\ \gdef \Geo {\mathrm{Geo}}\\ \gdef \N {\mathrm{N}} \\ \gdef \LN {\mathrm{LN}} \\ \gdef \U {\mathrm{U}} \\ \gdef \t {\mathrm{t}} \\ \gdef \F {\mathrm{F}} \\ \gdef \Exp {\mathrm{Exp}} \\ \gdef \Ga {\mathrm{Ga}} \\ \gdef \Be {\mathrm{Be}} \\ \gdef \NB {\mathrm{NB}}$
(再掲)
$X \sim \sum_{k=1}^K \pi_k \N(\mu_k, \sigma_k^2)$である時、その確率密度関数$f(x)$は、
$$\begin{aligned} f(x) = \sum_{k=1}^K \pi_k \cdot \dfrac{1}{\sqrt{2 \pi \sigma_k^2}} \exp[- \dfrac{(x – \mu_k)^2}{2 \sigma_k^2}] ~~ (- \infty \lt x \lt \infty) \end{aligned}$$であり、その特性値は、
$$\begin{aligned} &{\small ①平均:}E[X] = \sum_{k=1}^K \pi_k \mu_k \\ &{\small ②分散:}V[X] = \sum_{k=1}^K \pi_k ( \mu_k^2 + \sigma_k^2) – (\sum_{k=1}^K \pi_k \mu_k)^2 \\ &{\small ③積率母関数:}M^X(\theta) = \sum_{k=1}^K \pi_k \cdot \exp [ (\mu_k \theta + \dfrac{1}{2}\sigma_k^2 \theta^2)] \end{aligned}$$となる。
(①平均$E[X]$の証明)
$$\begin{aligned} E[X] &= \int_{-\infty}^{\infty} x f(x) dx \\ &= \sum_{k=1}^K \pi_k \cdot \underbrace{\int_{-\infty}^{\infty} x \cdot \dfrac{1}{\sqrt{2 \pi \sigma_k^2}} \exp[- \dfrac{(x – \mu_k)^2}{2 \sigma_k^2}]}_{\bullet} \\ &= \sum_{k=1}^K \pi_k \mu_k \\ &{\scriptsize (Y \sim \N(\mu_k, \sigma_k^2)とした時、\bulletはE[Y]に相当)} \end{aligned}$$
(②分散$V[X]$の証明)
$V[X]$を求めるにあたり、まず$E[X^2]$を求める。
$$\begin{aligned} E[X^2] &= \int_{-\infty}^{\infty} x^2 f(x) dx \\ &= \sum_{k=1}^K \pi_k \cdot \underbrace{\int_{-\infty}^{\infty} x^2 \cdot \dfrac{1}{\sqrt{2 \pi \sigma_k^2}} \exp[- \dfrac{(x – \mu_k)^2}{2 \sigma_k^2}]}_{\circ} \\ &= \sum_{k=1}^K \pi_k \cdot (\sigma_k^2 + \mu_k^2) \\ &{\scriptsize (Y \sim \N(\mu_k, \sigma_k^2)とした時、\circはE[Y^2](=V[Y] + (E[Y])^2)に相当)} \end{aligned}$$
よって、
$$\begin{aligned} V[X] &= E[X^2] – (E[X])^2 \\ &= \sum_{k=1}^K \pi_k ( \mu_k^2 + \sigma_k^2) – (\sum_{k=1}^K \pi_k \mu_k)^2 \end{aligned}$$
(③積率母関数$M^X(\theta)$の証明)
$$\begin{aligned} M^X(\theta) &= E[e^{\theta X}] \\ &= \int_{-\infty}^{\infty} e^{\theta x} \cdot \sum_{k=1}^K \pi_k \cdot \dfrac{1}{\sqrt{2 \pi \sigma_k^2}} \exp[- \dfrac{(x – \mu_k)^2}{2 \sigma_k^2}] \\ &= \sum_{k=1}^K \pi_k \underbrace{\int_{-\infty}^{\infty} e^{\theta x} \dfrac{1}{\sqrt{2 \pi \sigma_k^2}} \exp[- \dfrac{(x – \mu_k)^2}{2 \sigma_k^2}]}_{\triangle} \\ &= \sum_{k=1}^K \pi_k \cdot \exp [ (\mu_k \theta + \dfrac{1}{2}\sigma_k^2 \theta^2)] \\ &{\scriptsize (Y \sim \N(\mu_k, \sigma_k^2)とした時、\triangleはM^Y(\theta)に相当。参照:<連続分布>:「2. (単変量)正規分布」)} \end{aligned}$$
(参照:<連続分布>:「2. (単変量)正規分布」)