c*****l 发帖数: 631 | 1 marketing行业工作了两年,做的主要是campaign level的分析,最近开始搞customer
level的churn modeling and survival modeling,感觉有意思多了,另外手头还在搞数
据挖掘,用的工具是Angoss公司的Knowledge studio plus,本来喜欢SAS enterprise
miner, 但公司觉得太贵,说老实话,SAS公司的确太黑了,每年的liscenc超贵,尤其
对企业。所以将就着用这个studio搞决策树分析,功能上还凑合,请问这里有搞这方面的同仁吗?聊
聊看。 |
s*****n 发帖数: 2174 | |
q**j 发帖数: 10612 | 3 datamining和普通的regression etc有啥不一样的呢? |
d**n 发帖数: 198 | 4 weka
customer
面的同仁吗?聊
【在 c*****l 的大作中提到】![](/moin_static193/solenoid/img/up.png) : marketing行业工作了两年,做的主要是campaign level的分析,最近开始搞customer : level的churn modeling and survival modeling,感觉有意思多了,另外手头还在搞数 : 据挖掘,用的工具是Angoss公司的Knowledge studio plus,本来喜欢SAS enterprise : miner, 但公司觉得太贵,说老实话,SAS公司的确太黑了,每年的liscenc超贵,尤其 : 对企业。所以将就着用这个studio搞决策树分析,功能上还凑合,请问这里有搞这方面的同仁吗?聊 : 聊看。
|
o****o 发帖数: 8077 | |
S******y 发帖数: 1123 | 6 I used SAS enterprise miner before..
I do not like it...
many times it could not find a proper tree..
the user interface is stone-age. (looks like development of SAS/AF) |
s***x 发帖数: 293 | 7
That is because you don't know how to use it correctly!
【在 S******y 的大作中提到】![](/moin_static193/solenoid/img/up.png) : I used SAS enterprise miner before.. : I do not like it... : many times it could not find a proper tree.. : the user interface is stone-age. (looks like development of SAS/AF)
|
S******y 发帖数: 1123 | 8 Thanks. No worry.
I have switched to R tree package :-) |
b*******y 发帖数: 30 | 9 CART is pretty good, and relatively inexpensive |
b*******y 发帖数: 30 | 10 CART is pretty good, and relatively inexpensive |
s*********e 发帖数: 1051 | 11 i think SAS eminer is a good tool if used correctly.
i like to write code to run eminer on unix and find the segmentation before
building models.
an interesting function of eminer is that it can generate recoding programs
(such as scoring code... ), which is pretty neat. |