167.txt

This complete matlab for neural network

发信人: yaomc (白头翁&山东大汉), 信区: DataMining
标  题: Some useful thinking about DM to one DBMiner.
发信站: 南京大学小百合站 (Fri Mar 29 15:03:40 2002), 站内信件

Written by Qian Weining, I only post it here.

There are quite a lot of reviews for data mining softwares.
Some can be found in www.kdnuggets.com.
And I know that in the book:
"Building Data Mining Applications for CRM"
there is one section dicsussing the problem in quite details.
The Chinese edition of the book, translated by my lab-mates,
has been published by 'People's Post & Tele-comm Publisher'.
The authors of that book used to write one book named:
Data Warehousing, Data Mining & OLAP
around 1997, which is photo-copied by 世界图书出版公司.
Although the book is quite 'old',
the methods introduced in it is quite classic,
and the book is easy to find in Beijing or Shanghai.
And, I think the book is written for managers.

According to statistics you mentioned,
my opinion is that:

data mining =\= statistics
data mining =\= machine learning
data mining =\= pattern recognition
text mining =\= information retrieval
text mining =\= natural language processing
web mining =\= text mining + image mining
web mining =\= web log mining

I found from the web site of some members in this email group that
these concepts are confused，maybe led by some other people.
They are related, since some technologies are shared, or some are the
descendants of some others.
Since data mining is hot, so many people from other fields try to keep
close to it, for it is easy to get more profit, fund, customers,
papers, etc..
We must keep in mind that data mining applications are for managers
while some other applications are for statisticians, engineers, or 
some
other professionals. The data data mining applications to process is
always complex: large-volumned, missing-valued, mix-typed,
high-dimensional, full-of-noises, inconsistent, etc..

I am not familiar with management science.
In my idea, large chance to be wrong :),
the enterprises in China haven't organized their data well in 
databases
or data warehouses,
and they may not have listed their requests. Therefore, it is hard to
do mining for them. And, they may not feel competetive yet. :-)  

According to the data mining tools, I think 'general-purpose' data
mining tools has no future. :-) Different with other fields in IT, 
data
mining has no unified theoretical background or unified application
background. Therefore, find something general is impossible. :-) For
specific application, specific technology may be useful. However, the
approach Microsoft taken is promising. hehe

According to myself. I only used to attend some conferences in Hong
Kong. I have never studied there, although I am in Hong Kong now. I
think you must take me for another guy. :))

best,
-weining


--

Welcome to http://datamining.bbs.lilybbs.net.

※ 来源:．南京大学小百合站 bbs.nju.edu.cn．[FROM: 202.204.36.15]