These are chat archives for JohnsonHsieh/DSC2016-R

15th
Jul 2016
clariceliu
@clariceliu
Jul 15 2016 01:36

Wush Wu
@wush978
Jul 15 2016 01:48

`is.nan`
clariceliu
@clariceliu
Jul 15 2016 01:48

gscdelia
@gscdelia
Jul 15 2016 02:29

gscdelia
@gscdelia
Jul 15 2016 02:59

thinktwice812
@thinktwice812
Jul 15 2016 03:08

thinktwice812
@thinktwice812
Jul 15 2016 03:23

clariceliu
@clariceliu
Jul 15 2016 03:27
slide 34, Supp應該是60%?
Johnson
@JohnsonHsieh
Jul 15 2016 03:29
@clariceliu Thx and fixed
@thinktwice812 thx for the info :)
azephyrr
@azephyrr
Jul 15 2016 04:15

XC
@xchehub
Jul 15 2016 04:20
rules2.fulltime <- {} 是不是要改用 rules2.fulltime <- local({}) ?
Johnson
@JohnsonHsieh
Jul 15 2016 04:21
@azephyrr , RE lift值，只能說越大越好...個人看到 lift > 1.3 會滿開心的
@XC 其實都可以，省略local的原因是我的標準答案忘了check local({}) ...
azephyrr
@azephyrr
Jul 15 2016 04:27
got you! thanks!
Johnson
@JohnsonHsieh
Jul 15 2016 05:49
@gscdelia RE: 相似度，我下午會提到
@gscdelia RE: 均方差, (mean squared error, MSE)
Ninerh
@Ninerh
Jul 15 2016 06:26

clariceliu
@clariceliu
Jul 15 2016 06:29
Shapiro–Wilk test 是這個嗎?
planetoid
@planetoid
Jul 15 2016 06:36
@Ninerh: 按 skip() 可以看到解答
Ninerh
@Ninerh
Jul 15 2016 06:37
@planetoid : thank you :smile:
planetoid
@planetoid
Jul 15 2016 06:37

Johnson
@JohnsonHsieh
Jul 15 2016 06:37
Ninerh
@Ninerh
Jul 15 2016 06:38
@planetoid : 可以看 y 軸的 height 值，愈小愈近
planetoid
@planetoid
Jul 15 2016 06:38
@Ninerh : 謝謝
planetoid
@planetoid
Jul 15 2016 06:49

there is no package called ‘mclust’ 錯誤訊息 @@
emiliytw
@emiliytw
Jul 15 2016 06:52

planetoid
@planetoid
Jul 15 2016 06:53

emiliytw
@emiliytw
Jul 15 2016 06:55

SUCKCOOL
@ZiTUNG
Jul 15 2016 07:01

Johnson
@JohnsonHsieh
Jul 15 2016 07:03
factor analysis 是縮減 「欄位」
cluster analysis 是縮減 「觀察值」
SUCKCOOL
@ZiTUNG
Jul 15 2016 07:04

ianlee0609
@ianlee0609
Jul 15 2016 07:53
plotcluster(shapes\$x, cl2)

clariceliu
@clariceliu
Jul 15 2016 09:01

gscdelia
@gscdelia
Jul 15 2016 09:03

clariceliu
@clariceliu
Jul 15 2016 09:04

gscdelia
@gscdelia
Jul 15 2016 14:12

Johnson
@JohnsonHsieh
Jul 15 2016 16:41