How to make the normality ?

送交者: cornbug 于 2012-03-25, 15:38:02:

Suppose the variables are not normally distributed, how to transform them to achieve normality ? Will an appropriate transformation be acceptable in general?

所有跟贴:

你想干什么？ (无内容) - 短江学者 (0 bytes) 2012-03-25, 17:16:21 (655439)
- 这问话的口气好像玉米虫要图谋不轨似的 (无内容) - 通知 (0 bytes) 2012-03-25, 18:04:51 (655444)
  - 这适合你, 今天的 - cornbug (83 bytes) 2012-03-25, 20:10:15 (655447)
    - 谢了，真够漂亮 (无内容) - 通知 (0 bytes) 2012-03-26, 00:27:34 (655477)
    - 这张太色。 (无内容) - 短江学者 (0 bytes) 2012-03-25, 20:26:01 (655453)
      - 桃不色，是运 (无内容) - cornbug (0 bytes) 2012-03-25, 20:35:29 (655455)
- Thanks downstairs. - cornbug (147 bytes) 2012-03-25, 17:55:51 (655442)
先把X转为U(uniform)，再Y= probit function(U)，则Y~N(0,1) - 羽矢 (364 bytes) 2012-03-25, 16:24:36 (655436)
- The other question - cornbug (85 bytes) 2012-03-25, 18:00:37 (655443)
  - 不懂得 (无内容) - 羽矢 (0 bytes) 2012-03-25, 18:26:56 (655445)
such transformations all depend on assumption - conner (171 bytes) 2012-03-25, 16:03:29 (655435)

加跟贴