q**j
发帖数: 10612 | 1 今天又看了看Numpy，很好很强大。和matlab,R这些的差别很小了已经。 |
h******a
发帖数: 198 | |
q**j
发帖数: 10612 | 3 statpy, scipy and numpy
looks people are developing all kinds of tools for python
【在 h******a 的大作中提到】
: python也可以做统计分析啊？
|
c********r
发帖数: 520 | |
q**j
发帖数: 10612 | 5 any reason? numpy is written based on lapack, same as matlab. so numpy
should be much faster than R, and may suffer less from the memory
restriction. |
s*****n
发帖数: 2174 | 6 Then, you should as yourself a question: Why didn't Matlab take over R, give
n so many advantages?
Most people use R as a stat tool, not a generic tool for scientific computat
ion. As long as the stat package in Matlab, NumPy, etc are not as rich as in
R, R is probably still prefered.
【在 q**j 的大作中提到】
: any reason? numpy is written based on lapack, same as matlab. so numpy
: should be much faster than R, and may suffer less from the memory
: restriction.
|
q**j
发帖数: 10612 | 7 my thinking is this:
matlab is a corporation focused on profit and it has limited resources. so
the company will try to maximize its profit, which means that they won't
devote their time and effort in the way R people like for quite a while.
however, freelancers in Python is just like freelancers in R. They are
motivated and have tons of expertise and time to develop stuff. in fact, i
know Python because of a friend of mine is using it now to run regressions
everyday.
i have not tried python ext
【在 s*****n 的大作中提到】
: Then, you should as yourself a question: Why didn't Matlab take over R, give
: n so many advantages?
: Most people use R as a stat tool, not a generic tool for scientific computat
: ion. As long as the stat package in Matlab, NumPy, etc are not as rich as in
: R, R is probably still prefered.
|
s*****n
发帖数: 2174 | 8 Well, Python (or Matlab) can do statistics, but the scope is pretty limited.
There are indeed lots of Python enthusiatists, but do you think they will s
pend time to engineer stat models into Python? I think it is less likely. Fo
r new methodologies published 3 months ago, you can probably find R packages
for it. This just doesn't happen on Python, Matlab, even SAS.
For the advantages you mentioned, such as speed and memory, they are not usu
al concerns among R users. I had used R for 5 years in
【在 q**j 的大作中提到】
: my thinking is this:
: matlab is a corporation focused on profit and it has limited resources. so
: the company will try to maximize its profit, which means that they won't
: devote their time and effort in the way R people like for quite a while.
: however, freelancers in Python is just like freelancers in R. They are
: motivated and have tons of expertise and time to develop stuff. in fact, i
: know Python because of a friend of mine is using it now to run regressions
: everyday.
: i have not tried python ext
|
d******e
发帖数: 7844 | 9 针对不同人群。R是免费的，各有所需，不能一概而论。
【在 q**j 的大作中提到】
: my thinking is this:
: matlab is a corporation focused on profit and it has limited resources. so
: the company will try to maximize its profit, which means that they won't
: devote their time and effort in the way R people like for quite a while.
: however, freelancers in Python is just like freelancers in R. They are
: motivated and have tons of expertise and time to develop stuff. in fact, i
: know Python because of a friend of mine is using it now to run regressions
: everyday.
: i have not tried python ext
|
q**j
发帖数: 10612 | 10 python也是不要钱的。各个领域不同。我干的活60%不能用R，due to size
limitations. otherwise, i have no problem with learning R only. |
|
|
P****D
发帖数: 11146 | 11 S-PLUS may.
【在 c********r 的大作中提到】
: R 不会的
|
h******a
发帖数: 198 | 12 R 里面也不少用LINPACK写的，比如princomp，做PCA的。 |
o****o
发帖数: 8077 | 13 can u manipulate your data in DB first? then use R on the FINAL table?
【在 q**j 的大作中提到】
: python也是不要钱的。各个领域不同。我干的活60%不能用R，due to size
: limitations. otherwise, i have no problem with learning R only.
|
N**D
发帖数: 10322 | 14 R is only live for stat community
no company wants analysis in one platform, and production on another. waste
of resource.
limited.
s
Fo
packages
usu
even
explor
pretty
But
【在 s*****n 的大作中提到】
: Well, Python (or Matlab) can do statistics, but the scope is pretty limited.
: There are indeed lots of Python enthusiatists, but do you think they will s
: pend time to engineer stat models into Python? I think it is less likely. Fo
: r new methodologies published 3 months ago, you can probably find R packages
: for it. This just doesn't happen on Python, Matlab, even SAS.
: For the advantages you mentioned, such as speed and memory, they are not usu
: al concerns among R users. I had used R for 5 years in
|
N**D
发帖数: 10322 | 15 It is the R that has the memory issue
【在 h******a 的大作中提到】
: R 里面也不少用LINPACK写的，比如princomp，做PCA的。
|
q**j
发帖数: 10612 | 16 DB has limited functions. not handy.
【在 o****o 的大作中提到】
: can u manipulate your data in DB first? then use R on the FINAL table?
|
