用什么方法当然决定于什么误差怎么分布

送交者: wasqaz 于 2009-05-30, 16:49:37:

回答: Gauss-Markov定理并不要求误差为高斯分布。由 wasguru 于 2009-05-30, 16:35:33:

不是数学简单。
你倒过来想也行，开始的时候这种方法简单，后来有人指出这种方法对应的误差分布是什么。

你看一下the errors have expectation zero and are uncorrelated and have equal variances，是不是等价于高斯分布？

In a linear model in which the errors have expectation zero and are uncorrelated and have equal variances, a best linear unbiased estimator (BLUE) of the coefficients is given by the least-squares estimator.

http://en.wikipedia.org/wiki/Gauss%E2%80%93Markov_theorem

所有跟贴:

当然不是数学简单这么简单 - wasguru (50 bytes) 2009-05-30, 17:00:48 (331338)
- 跟他开玩笑没看出来？ - wasqaz (15 bytes) 2009-05-30, 17:02:51 (331340)
你就不会多读一句？ - wasguru (23 bytes) 2009-05-30, 16:54:02 (331334)
- 这算我没想到，不过不影响论点 - wasqaz (18 bytes) 2009-05-30, 16:57:41 (331336)
  - the errors have expectation zero and are uncorrelated and have equal variances - wasqaz (113 bytes) 2009-05-30, 17:01:37 (331339)
    - 这个条件连非对称r.v..都能满足,怎么会和高斯分布等价? (无内容) - LaoXiao (0 bytes) 2009-05-30, 18:07:06 (331352)
      - 是么？<Xi>=0, <Xi Xj>=0 - wasqaz (90 bytes) 2009-05-30, 18:25:56 (331353)
        
        Let {Xi} be any i.i.d. r.v. with finite 2nd moment, then ...... - LaoXiao (92 bytes) 2009-05-30, 20:50:43 (331369)
        
        有何不可？ - wasguru (100 bytes) 2009-05-30, 18:34:34 (331354)
        
        方差相等也是两个随机变量之间的关系 - wasguru (114 bytes) 2009-05-30, 18:53:11 (331359)
        
        and have equal variances - wasqaz (25 bytes) 2009-05-30, 18:49:01 (331356)
    - 等价? 差远了去. (无内容) - LaoXiao (0 bytes) 2009-05-30, 17:58:08 (331351)
    - 当然不等价 (无内容) - wasguru (0 bytes) 2009-05-30, 17:04:08 (331341)

加跟贴