由买买提看人间百态

boards

本页内容为未名空间相应帖子的节选和存档,一周内的贴子最多显示50字,超过一周显示500字 访问原贴
Statistics版 - 关于统计研究生选修计算机系课的建议
相关主题
有人有 %GNBC 吗?training dataset validation dataset and test dataset
今天SAS BASE 通过关于decision tree
SAS neural network 和 SVM 的macro求两本电子书-包子答谢
一道面试题,向本版求教一下。Ebook-The elements of statistical learning:data mining,inference,and prediction.2nd edition(2009)
Re: 请推荐一本data mining入门教材?谈谈最近两次面试经历
Re: 请推荐nonparametric regression 的入门经典书book for data mining or predictive modeling
need a good book on Statistical Modelingmodel selection一般都用什么方法
good classification methods for high dimension data请推荐一本学习Data Mining 的书, 谢谢。
相关话题的讨论汇总
话题: data话题: c++话题: learning话题: sas
进入Statistics版参与讨论
1 (共1页)
c*****l
发帖数: 297
1
关于统计研究生选修计算机系课的建议
1: Programming level
C/C++ (Two semesters)
Data Structure
An introduction to algorithm
2: Senior Master Level
Data Base Management (SQL or Oracle)
Machine Learning or Statistics Learning (Classifiers or prediction
algorithms)
3: Senior PhD student Level
Programming Language (lexical analysis, syntax analysis, semantic
analysis, Object Programming)
Operating System (Linux or others)
Artificial Intelligence
如果你还有时间选修计算机的课,那么你就可以并修一个计算机的硕士了。
x******0
发帖数: 1025
2
加精加精!!
G***s
发帖数: 10030
3
baozi~
正好回答了我的问题,不过学了这些,找工作性质是不是有偏离统计了?
O*O
发帖数: 2284
4
出路多了不少
我的体会是下面5个方面会3个就很好了
1. Database (Oracle, MySQL, SQL Server)
2. Script (R, Matlab, Perl, Python)
3. Web Interface/Application (Javascript, .Net, VC#)
4. SAS, Excel
5. C++, C#

【在 G***s 的大作中提到】
: baozi~
: 正好回答了我的问题,不过学了这些,找工作性质是不是有偏离统计了?

c*****l
发帖数: 297
5
修好C++其实就是为R,SAS,PERL服务的。当C++, Data Structure修完后, R SAS应该不难
了。
C#一般很少在统计中用到。PERL在BIOINFORMATICS用的多。

【在 O*O 的大作中提到】
: 出路多了不少
: 我的体会是下面5个方面会3个就很好了
: 1. Database (Oracle, MySQL, SQL Server)
: 2. Script (R, Matlab, Perl, Python)
: 3. Web Interface/Application (Javascript, .Net, VC#)
: 4. SAS, Excel
: 5. C++, C#

w********3
发帖数: 1503
6
mark!
c****r
发帖数: 576
7
说的不错,相对于C/C++,什么R/Matlab之类都是浮云,做做研究还行,做工具就差远
了。

【在 c*****l 的大作中提到】
: 关于统计研究生选修计算机系课的建议
: 1: Programming level
: C/C++ (Two semesters)
: Data Structure
: An introduction to algorithm
: 2: Senior Master Level
: Data Base Management (SQL or Oracle)
: Machine Learning or Statistics Learning (Classifiers or prediction
: algorithms)
: 3: Senior PhD student Level

d********I
发帖数: 9
8
恩,看一些机器学习的书挺有用的。
推荐The Elements of Statistical Learning, Hastie et al., 2009
http://www.amazon.com/Elements-Statistical-Learning-Prediction-
A*******s
发帖数: 3942
9
sometimes it really depends on the company.
Our company is more and more restrict on IT risk management. Right now i am
not allowed to install any free software in my laptop. Unix servers have
very limited permission for users. At the beginning I was surprised to see
some people using PC SAS to handle some large datasets. Now I get used to it
.
If the company doesn't care about efficiency, why should we care.

不难

【在 c*****l 的大作中提到】
: 修好C++其实就是为R,SAS,PERL服务的。当C++, Data Structure修完后, R SAS应该不难
: 了。
: C#一般很少在统计中用到。PERL在BIOINFORMATICS用的多。

D******n
发帖数: 2836
10
回错贴?
unix server一定要配一个admin 否则就会乱套。所以成本还是蛮高的。不过以前我们
学校都是一个admin管理整栋大楼,这个费用也可以忽略了。反倒是现在公司,没有
admin,或者说非常驻的,非常麻烦。装个软件都要找组里的一个人,这个人再跟另外
一些人谈,omg。问题是组里这个人自己也不是很懂,还老问我为啥要装,老气横秋。
effieicncy啊efficiency。

am
it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

相关主题
Re: 请推荐nonparametric regression 的入门经典书training dataset validation dataset and test dataset
need a good book on Statistical Modeling关于decision tree
good classification methods for high dimension data求两本电子书-包子答谢
进入Statistics版参与讨论
A*******s
发帖数: 3942
11
my point is that strong skills on CS might be useless in some companies
because of stupid bureaucracy.
last week i was blamed by the MIS team because my pc submitted a query when
one database was being loaded... WTF...

【在 D******n 的大作中提到】
: 回错贴?
: unix server一定要配一个admin 否则就会乱套。所以成本还是蛮高的。不过以前我们
: 学校都是一个admin管理整栋大楼,这个费用也可以忽略了。反倒是现在公司,没有
: admin,或者说非常驻的,非常麻烦。装个软件都要找组里的一个人,这个人再跟另外
: 一些人谈,omg。问题是组里这个人自己也不是很懂,还老问我为啥要装,老气横秋。
: effieicncy啊efficiency。
:
: am
: it

x**********0
发帖数: 163
12
LZ 你可以给我留个联络方式吗?我很想以后多跟你请教一些东西,谢谢
s******r
发帖数: 1524
13
I would be surprised that your company would allow you install free software
. Most companies say No to that. They also block personal email. The most
important for them is the safety not efficiency.

am
it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

A*******s
发帖数: 3942
14
Not just free software but any software. You need to ask IT guys to install
software remotely for you. But those guys know nothing about configuration!
Some of them don't even know what are SAS and Revo R... They tried several
times and failed, and then gave me temporary admin right to install them...

software

【在 s******r 的大作中提到】
: I would be surprised that your company would allow you install free software
: . Most companies say No to that. They also block personal email. The most
: important for them is the safety not efficiency.
:
: am
: it

f****r
发帖数: 1140
15
I guess the data is not large enough.
maybe at summarized level?

am
it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

f****r
发帖数: 1140
16
This is true for many large companies.
Especially for banks, i guess. they care much about information security.

install
!

【在 A*******s 的大作中提到】
: Not just free software but any software. You need to ask IT guys to install
: software remotely for you. But those guys know nothing about configuration!
: Some of them don't even know what are SAS and Revo R... They tried several
: times and failed, and then gave me temporary admin right to install them...
:
: software

f****r
发帖数: 1140
17
Nice post..
It also depends on what kind of career you like to choose.
For those tech/data based positions, Data Base Management is a very useful
skill.

【在 c*****l 的大作中提到】
: 关于统计研究生选修计算机系课的建议
: 1: Programming level
: C/C++ (Two semesters)
: Data Structure
: An introduction to algorithm
: 2: Senior Master Level
: Data Base Management (SQL or Oracle)
: Machine Learning or Statistics Learning (Classifiers or prediction
: algorithms)
: 3: Senior PhD student Level

s******r
发帖数: 1524
18
It happens all the time. I am so surprised they would gave you temporary
admin. Actually it is the policy of the company, specially for financial
institute , like bank. Very very few companies would allow you control the
computer. You would be used to it sooner or later.

install
!

【在 A*******s 的大作中提到】
: Not just free software but any software. You need to ask IT guys to install
: software remotely for you. But those guys know nothing about configuration!
: Some of them don't even know what are SAS and Revo R... They tried several
: times and failed, and then gave me temporary admin right to install them...
:
: software

c*****l
发帖数: 297
19
你在回答什么问题?别把楼整歪了

now i am
have
to see
used to it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

m****0
发帖数: 253
20
纯属瞎说、一派胡言

不难

【在 c*****l 的大作中提到】
: 修好C++其实就是为R,SAS,PERL服务的。当C++, Data Structure修完后, R SAS应该不难
: 了。
: C#一般很少在统计中用到。PERL在BIOINFORMATICS用的多。

相关主题
Ebook-The elements of statistical learning:data mining,inference,and prediction.2nd edition(2009)model selection一般都用什么方法
谈谈最近两次面试经历请推荐一本学习Data Mining 的书, 谢谢。
book for data mining or predictive modeling230 Variables and 4400 Observations 算是high-dimensional data么
进入Statistics版参与讨论
A*******s
发帖数: 3942
21
我就是说,要是只搞统计的话,有不少大公司可能压根就没地方让你用这些技能。

【在 c*****l 的大作中提到】
: 你在回答什么问题?别把楼整歪了
:
: now i am
: have
: to see
: used to it

f********t
发帖数: 117
22
in my opinion, introductory C/C++(at most + C++ in data structure) is enough
to get you familiar with common syntax, control structure, data types in
programming languages. the you learn scripting languages like perl, python
or php by yourself, because they aren't taught anywhere, but they are
required for data analyst positions, esp perl. typically a data analyst
could work in Linux environment. so introduction to UNIX course will be
helpful.
typical database course in MS program are a little more theoretical than
what you need. you learned a lot about db design, how table are normalized
, or other shit I forgot. but for a data analyst, all you need to know is
sql commands. you will never need to design a db, or even tables. so I would
say take sql courses in community colleges should be more helpful for
getting a job. some people use sas everyday, then that is all they need to
know.
for courses like machine learning, it is more specialized. you dont need to
take the course, unless you want do machine learning.
c*****l
发帖数: 297
23
关于统计研究生选修计算机系课的建议
1: Programming level
C/C++ (Two semesters)
Data Structure
An introduction to algorithm
2: Senior Master Level
Data Base Management (SQL or Oracle)
Machine Learning or Statistics Learning (Classifiers or prediction
algorithms)
3: Senior PhD student Level
Programming Language (lexical analysis, syntax analysis, semantic
analysis, Object Programming)
Operating System (Linux or others)
Advanced Algorithm
Artificial Intelligence
如果你还有时间选修计算机的课,那么你就可以并修一个计算机的硕士了。
所有的这些计算的课都是普通的基础课,一般编程的课都修2个学期,计算机系的本科
生都是修2个学期的C++,基础一定要打稳。当然不一定要你把C++搞透,那是不可能的。至少在
1000Line内的CODE你能看懂。别搞那么复杂,修了2个学期的C++应该能看懂C++ STL 和基本
语句吧。其实修这个课的途中,老师布置的作业是个很好的训练过程。可以多和本科生交流,多
问。一定要练才行,刚开始是模拟别人的例子写,能看懂别人在干什么。慢慢得你就上路了。
我的计算机的本科生课都是在这边上的,刚开始修得很痛苦,甚至在国内的TAOBAO上买
相应的本科生视频学习。国内的很多大学都把这些基础课录制下来,大家可以去找相应的视频学习
,有些老师讲得很好。
等把编程的基础过关后,DATABASE 和MACHINE LEARNING对找工作是挺有用的。
比如过统计系修的GENERALIZED LINEAR MODEL, 或者是Generalized Linear and Mix
Model,或者说EXPERIMENT DESIGN 对于MACHINE Learning来说都是一个小例子。
Machine Learning里面有很多好的方法。比如说
K Nearest Neighbor,
Support Vector Machine,
Neural Network,
Booting and Bagging,
Random Forest,
Probabilistic Graphic Model ( Naive Bayes Model, Bayesian Network
Model)
Decision Tree Model: ID3, J48, Classification and Regression Tree
.......
这里我就不一一举例
DATABASE 的好处 我就不用说大家都知道,统计工作者能能够Handle大型数据是挺好的
,尤其的大的公司都是GB甚至TB的数据。不管去学术界也好,去研究所工作也好,还是去公司里工
作也好,单凭会几个简单的Model,甚至懂几个SAS语句,一般情况下很难找到工作。
希望还在挣扎的同行们,咬牙挺过这些艰难的日子。
x******0
发帖数: 1025
24
加精加精!!
G***s
发帖数: 10030
25
baozi~
正好回答了我的问题,不过学了这些,找工作性质是不是有偏离统计了?
O*O
发帖数: 2284
26
出路多了不少
我的体会是下面5个方面会3个就很好了
1. Database (Oracle, MySQL, SQL Server)
2. Script (R, Matlab, Perl, Python)
3. Web Interface/Application (Javascript, .Net, VC#)
4. SAS, Excel
5. C++, C#

【在 G***s 的大作中提到】
: baozi~
: 正好回答了我的问题,不过学了这些,找工作性质是不是有偏离统计了?

c*****l
发帖数: 297
27
修好C++其实就是为R,SAS,PERL服务的。当C++, Data Structure修完后, R SAS应该不难
了。
C#一般很少在统计中用到。PERL在BIOINFORMATICS用的多。

【在 O*O 的大作中提到】
: 出路多了不少
: 我的体会是下面5个方面会3个就很好了
: 1. Database (Oracle, MySQL, SQL Server)
: 2. Script (R, Matlab, Perl, Python)
: 3. Web Interface/Application (Javascript, .Net, VC#)
: 4. SAS, Excel
: 5. C++, C#

w********3
发帖数: 1503
28
mark!
c****r
发帖数: 576
29
说的不错,相对于C/C++,什么R/Matlab之类都是浮云,做做研究还行,做工具就差远
了。

【在 c*****l 的大作中提到】
: 关于统计研究生选修计算机系课的建议
: 1: Programming level
: C/C++ (Two semesters)
: Data Structure
: An introduction to algorithm
: 2: Senior Master Level
: Data Base Management (SQL or Oracle)
: Machine Learning or Statistics Learning (Classifiers or prediction
: algorithms)
: 3: Senior PhD student Level

d********I
发帖数: 9
30
恩,看一些机器学习的书挺有用的。
推荐The Elements of Statistical Learning, Hastie et al., 2009
http://www.amazon.com/Elements-Statistical-Learning-Prediction-
相关主题
新手请教一个分类问题今天SAS BASE 通过
搞统计的人的怨念...SAS neural network 和 SVM 的macro
有人有 %GNBC 吗?一道面试题,向本版求教一下。
进入Statistics版参与讨论
A*******s
发帖数: 3942
31
sometimes it really depends on the company.
Our company is more and more restrict on IT risk management. Right now i am
not allowed to install any free software in my laptop. Unix servers have
very limited permission for users. At the beginning I was surprised to see
some people using PC SAS to handle some large datasets. Now I get used to it
.
If the company doesn't care about efficiency, why should we care.

不难

【在 c*****l 的大作中提到】
: 修好C++其实就是为R,SAS,PERL服务的。当C++, Data Structure修完后, R SAS应该不难
: 了。
: C#一般很少在统计中用到。PERL在BIOINFORMATICS用的多。

D******n
发帖数: 2836
32
回错贴?
unix server一定要配一个admin 否则就会乱套。所以成本还是蛮高的。不过以前我们
学校都是一个admin管理整栋大楼,这个费用也可以忽略了。反倒是现在公司,没有
admin,或者说非常驻的,非常麻烦。装个软件都要找组里的一个人,这个人再跟另外
一些人谈,omg。问题是组里这个人自己也不是很懂,还老问我为啥要装,老气横秋。
effieicncy啊efficiency。

am
it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

A*******s
发帖数: 3942
33
my point is that strong skills on CS might be useless in some companies
because of stupid bureaucracy.
last week i was blamed by the MIS team because my pc submitted a query when
one database was being loaded... WTF...

【在 D******n 的大作中提到】
: 回错贴?
: unix server一定要配一个admin 否则就会乱套。所以成本还是蛮高的。不过以前我们
: 学校都是一个admin管理整栋大楼,这个费用也可以忽略了。反倒是现在公司,没有
: admin,或者说非常驻的,非常麻烦。装个软件都要找组里的一个人,这个人再跟另外
: 一些人谈,omg。问题是组里这个人自己也不是很懂,还老问我为啥要装,老气横秋。
: effieicncy啊efficiency。
:
: am
: it

x**********0
发帖数: 163
34
LZ 你可以给我留个联络方式吗?我很想以后多跟你请教一些东西,谢谢
s******r
发帖数: 1524
35
I would be surprised that your company would allow you install free software
. Most companies say No to that. They also block personal email. The most
important for them is the safety not efficiency.

am
it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

A*******s
发帖数: 3942
36
Not just free software but any software. You need to ask IT guys to install
software remotely for you. But those guys know nothing about configuration!
Some of them don't even know what are SAS and Revo R... They tried several
times and failed, and then gave me temporary admin right to install them...

software

【在 s******r 的大作中提到】
: I would be surprised that your company would allow you install free software
: . Most companies say No to that. They also block personal email. The most
: important for them is the safety not efficiency.
:
: am
: it

f****r
发帖数: 1140
37
I guess the data is not large enough.
maybe at summarized level?

am
it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

f****r
发帖数: 1140
38
This is true for many large companies.
Especially for banks, i guess. they care much about information security.

install
!

【在 A*******s 的大作中提到】
: Not just free software but any software. You need to ask IT guys to install
: software remotely for you. But those guys know nothing about configuration!
: Some of them don't even know what are SAS and Revo R... They tried several
: times and failed, and then gave me temporary admin right to install them...
:
: software

f****r
发帖数: 1140
39
Nice post..
It also depends on what kind of career you like to choose.
For those tech/data based positions, Data Base Management is a very useful
skill.

【在 c*****l 的大作中提到】
: 关于统计研究生选修计算机系课的建议
: 1: Programming level
: C/C++ (Two semesters)
: Data Structure
: An introduction to algorithm
: 2: Senior Master Level
: Data Base Management (SQL or Oracle)
: Machine Learning or Statistics Learning (Classifiers or prediction
: algorithms)
: 3: Senior PhD student Level

s******r
发帖数: 1524
40
It happens all the time. I am so surprised they would gave you temporary
admin. Actually it is the policy of the company, specially for financial
institute , like bank. Very very few companies would allow you control the
computer. You would be used to it sooner or later.

install
!

【在 A*******s 的大作中提到】
: Not just free software but any software. You need to ask IT guys to install
: software remotely for you. But those guys know nothing about configuration!
: Some of them don't even know what are SAS and Revo R... They tried several
: times and failed, and then gave me temporary admin right to install them...
:
: software

相关主题
一道面试题,向本版求教一下。need a good book on Statistical Modeling
Re: 请推荐一本data mining入门教材?good classification methods for high dimension data
Re: 请推荐nonparametric regression 的入门经典书training dataset validation dataset and test dataset
进入Statistics版参与讨论
c*****l
发帖数: 297
41
你在回答什么问题?别把楼整歪了

now i am
have
to see
used to it

【在 A*******s 的大作中提到】
: sometimes it really depends on the company.
: Our company is more and more restrict on IT risk management. Right now i am
: not allowed to install any free software in my laptop. Unix servers have
: very limited permission for users. At the beginning I was surprised to see
: some people using PC SAS to handle some large datasets. Now I get used to it
: .
: If the company doesn't care about efficiency, why should we care.
:
: 不难

m****0
发帖数: 253
42
纯属瞎说、一派胡言

不难

【在 c*****l 的大作中提到】
: 修好C++其实就是为R,SAS,PERL服务的。当C++, Data Structure修完后, R SAS应该不难
: 了。
: C#一般很少在统计中用到。PERL在BIOINFORMATICS用的多。

A*******s
发帖数: 3942
43
我就是说,要是只搞统计的话,有不少大公司可能压根就没地方让你用这些技能。

【在 c*****l 的大作中提到】
: 你在回答什么问题?别把楼整歪了
:
: now i am
: have
: to see
: used to it

f********t
发帖数: 117
44
in my opinion, introductory C/C++(at most + C++ in data structure) is enough
to get you familiar with common syntax, control structure, data types in
programming languages. the you learn scripting languages like perl, python
or php by yourself, because they aren't taught anywhere, but they are
required for data analyst positions, esp perl. typically a data analyst
could work in Linux environment. so introduction to UNIX course will be
helpful.
typical database course in MS program are a little more theoretical than
what you need. you learned a lot about db design, how table are normalized
, or other shit I forgot. but for a data analyst, all you need to know is
sql commands. you will never need to design a db, or even tables. so I would
say take sql courses in community colleges should be more helpful for
getting a job. some people use sas everyday, then that is all they need to
know.
for courses like machine learning, it is more specialized. you dont need to
take the course, unless you want do machine learning.
m********l
发帖数: 791
45
mark~
1 (共1页)
进入Statistics版参与讨论
相关主题
请推荐一本学习Data Mining 的书, 谢谢。Re: 请推荐一本data mining入门教材?
230 Variables and 4400 Observations 算是high-dimensional data么Re: 请推荐nonparametric regression 的入门经典书
新手请教一个分类问题need a good book on Statistical Modeling
搞统计的人的怨念...good classification methods for high dimension data
有人有 %GNBC 吗?training dataset validation dataset and test dataset
今天SAS BASE 通过关于decision tree
SAS neural network 和 SVM 的macro求两本电子书-包子答谢
一道面试题,向本版求教一下。Ebook-The elements of statistical learning:data mining,inference,and prediction.2nd edition(2009)
相关话题的讨论汇总
话题: data话题: c++话题: learning话题: sas