6 Comments
访客 *lixiaoxu* @ 2009-02-25 14:46:21 写道:
这个问题文献中被引滥的成语是 if all you have is a hammer, everything looks like a nail
哈哈,所言极是!
——原帖发布于 2009-02-25 14:57:56
访客 *lixiaoxu* @ 2009-02-25 14:48:32 写道:
如有引用需要,可参考 http://en.wikipedia.org/wiki/Golden_hammer 条陈其出典
Guest *xiyting* @ 2009-02-25 15:12:09 originally posted:
Nonparametric statistics is distribution-free.
呃……谢谢,谢谢!貌似我经常把话写反……汗
——原帖发布于 2009-02-25 19:02:40
访客 *wind* @ 2009-02-25 19:13:51 写道:
不熟悉对各个分布进行检验。干脆画个图。直观一点。就我的实践而言,通常我宁可使用直接从数据本身产生的密度函数,而不是找一个接近的标准分布。主要因为对各种分布也不太熟悉。 😳
hist(len, breaks = 50, col = "blue", border = "blue",
freq = FALSE)
lines(density(len), col = "blue")
x = seq(1, 100, 1)
lines(x, dnorm(x, mean = mean(len), sd = sd(len)),
type = "l", col = "red")
lines(x, dpois(x, lambda = sd(len)), type = "l", col = "black")
legend(50, 0.07, c(" 数据密度", "正态分布", "泊松分布"),
text.col = c("blue", "red", "black"))
http://picasaweb.google.com/lh/photo/BdBUd_QCmXQRMenSn6ZOJA?feat=directlink
Sign in to join the discussion
Sign in with GitHub