일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |
- WBP
- dl
- 오블완
- 생성모델
- Stat110
- WinAPI
- deep learning
- NLP
- animation retargeting
- 폰트생성
- Generative Model
- 언리얼엔진
- 딥러닝
- Few-shot generation
- 디퓨전모델
- cv
- motion matching
- UE5
- GAN
- multimodal
- CNN
- Font Generation
- RNN
- ddpm
- BERT
- Unreal Engine
- userwidget
- 모션매칭
- Diffusion
- ue5.4
- Today
- Total
Deeper Learning
[Statistics 110] Lecture 30: Chi-Square, Student-t, Multivariate Normal 본문
[Statistics 110] Lecture 30: Chi-Square, Student-t, Multivariate Normal
Dlaiml 2023. 1. 28. 16:41Chi-Square Distribution(카이제곱 분포)
V=Z21+...+Z2n, Zj∼N(0,1)i.i.d일 때, V∼χ2(n).
χ2(1)은 Gamma(12,12)와 같은 분포, χ2는 Gamma(n2,12)와 같은 분포
t-Distribution(t-분포)
Z∼N(0,1)V∼χ2(n)T=Z√V/n∼tn
성질
- Symmetric: −T∼tn.
- n=1이면 코시분포를 따름
- n이 2이상이면 E(T)=0.
- 정규분포와 비슷하지만 더 heavy tailed(극단적인 값이 더 발생)
- n이 크면 표준정규분포와 매우 비슷함
n→inf을 큰 수의 법칙을 사용해서 증명해보면
T = \frac{Z}{\sqrt{V/n}} \sim t_n\\
에서 \frac {V_n}n은 1로 1의 확률로 수렴하므로 T \rightarrow Z도 1의 확률로 성립
Multivariate Normal(다변량 정규분포)
Random Vector \vec X=(X_1,...,X_n)에서 모든 t에 대한 선형 결합(t_1x_1+...+t_nx_n)이 정규분포를 따르면, 다변량 정규분포를 따른다.
예시) Z,W가 i.i.d N(0,1)일 때 (Z+2W,3Z+5W)는 다변량 정규분포를 따른다.
임의의 상수 s,t에 대해 선형 결합을 하면
s(Z+2W) + t(3Z+5W) = (s+3t)Z + (2s+5t)W \sim N
독립인 두 정규분포의 합도 정규분포임을 이전에 증명하였기 때문에 예시에 대한 증명완료
예시) Z는 표준정규분포, S는 random sign(-1 or 1) 서로 독립일 때, Z와 SZ는 표준정규분포를 따르지만 둘의 선형결합인 Z+SZ는 절반이 0인 분포로 정규분포를 따르지 않는다.
적률생성함수
다변량 정규분포를 따르는 확률변수 \vec X의 MGF는 E(e^ {\vec t\cdot \vec X}) .
E(e^ {\vec t\cdot \vec X}) =E(e^ {t_1\mu_1+...+t_k\mu_k+\frac 12Var(t_1X_1+...+t_kX_k)})
다변량 정규분포에서는 비상관관계(공분산이 0)이면 독립이다.

Reference
[0] https://www.youtube.com/playlist?list=PL2SOU6wwxB0uwwH80KTQ6ht66KWxbzTIo
'Statistics & Math > Statistics 110: Probability' 카테고리의 다른 글
[Statistics 110] Lecture 32: Markov Chains Continued (0) | 2023.01.28 |
---|---|
[Statistics 110] Lecture 31: Markov Chains (0) | 2023.01.28 |
[Statistics 110] Lecture 29: Law of Large Numbers and Central Limit Theorem (0) | 2023.01.28 |
[Statistics 110] Lecture 28: Inequalities (0) | 2023.01.28 |
[Statistics 110] Lecture 27: Conditional Expectation given an R.V. (0) | 2023.01.28 |