管理员
- 积分
- 20595
- 大米
- 颗
- 鳄梨
- 个
- 水井
- 尺
- 蓝莓
- 颗
- 萝卜
- 根
- 小米
- 粒
- 学分
- 个
- 注册时间
- 2011-3-12
- 最后登录
- 1970-1-1
|
本帖最后由 wwtpcsuper 于 2012-8-29 19:49 编辑
zwbdmm 发表于 2012-8-28 21:30 -baidu 1point3acres
在本科阶段的simulation和optimization, LA都接触过 convex mcmc svd, 怎么那些课程在研究生还要学么?
Give you an example.(actually a qualifier problem in our dept). 1point 3acres
Given x1 to xn follow poisson distribution, but the parameter lamda follows gamma distribution. While alpha is given and beta follows an inverse gamma distribution.
i) you have a gamma random number generator, please use gibbs sampler to generate x1 to xn
ii) you have a normal random number generator, please use metropolian-hasting algorithm to generate x1 to xn
If you think you will solve this problem easily, you are surely doing well in MCMC and do not need to learn master level computational statistics courses.
|
|