Why is Python so fashionable in machine learning? - page 2

New comment
 
Maxim Dmitrievsky:

P is not richer, all machinelerners work on python, P is used by statistics and other uneducated people such as local grief traders, because everything there is as simple as two fingers in three lines.

That's why there are so many libs and because every genius or student tends to do his or her own thing.

I think over time (a couple of years) Python will catch up with the number of R libraries.

 
forexman77:

It seems to me that over time (a couple of years) Python will catch up with the number of R libraries.

What do you mean? There are originally more of them. Then they rewrite to R

http://scikit-learn.org/stable/

start there.

If you want something cooler, use TensorFlow, Theano, PyTorch and others (if you need it :))

 
Maxim Dmitrievsky:

I've already answered you two times in different threads

You seem to forget everything after a while

that python is a high-level language that is very convenient for working with vectors, matrices and neural networks

the speed is not important there because the main time consuming operations are done on the pluses and on video cards

All that is slow is preprocessing and is done only once

Ok, let's assume I forgot python and don't remember about vector/matrix operations. Is it possible to do the following there (matlab code on the command line)

>> x= [1 2 3]

x =

     1     2     3

>> y = [4 5 6]

y =

     4     5     6

>> z = x+y % векторная операция, так на питоне можно?

z =

     5     7     9

***

 
Alexey Volchanskiy:

Okay, let's assume that I forgot about python and I don't remember about vector/matrix operations. Is it possible to do the following (matlab code on the command line)

***

Well, of course you can

 
Maxim Dmitrievsky:

What do you mean? There are more of them initially. Then they're rewritten to R.

http://scikit-learn.org/stable/

start there.

If you want something cooler, use TensorFlow, Theano, PyTorch and others (if you need it :))

Still demonstrating dull ignorance, just too lazy to google? Well, just not to look too ...

 
SanSanych Fomenko:

Continuing to show dull ignorance, just too lazy to google it? Well, just so you don't look too...

here comes R - monks have arrived, now they will start to lecture you about how wonderful he is, I counted on it hahahaha )) language for students

 
forexman77:

In Python there are plenty of examples and forums where you can ask if you don't understand something. In R you have to understand everything yourself and it takes a lot of time and I haven't seen any forums on R at all (except one subforum).

Plus now there is the NumPY library. Vector calculations are much faster. But, also noticed all the same, that the code in the R console, in my opinion, is faster.

Basically nothing has changed Python is much friendlier to learn and understand, and R is richer and has more stuff in it in terms of machine learning.

You just don't know about R, it's full of forums, it's well supported, it has a huge amount of useful literature for us both in the form of books and articles. I don't know anything like that in Python. The usual modern practice for formulating statistical thoughts is code in R, and quite rarely in Python.

If you take the statistics of the use of the two languages, the prevalence of them are about the same, but in Python there are a lot of users who write websites. If we evaluate usefulness of these two languages for trading, R has an undoubted advantage. Originally R (the paid prototype S of the 70's) was designed to solve exclusively statistical problems and was called so: the system of graphs and statistics.


One last thing.

What to pay attention to google. Take Microsoft. As of today R is part of Microsoft's software.

 
Alexey Volchanskiy:

Okay, let's assume that I forgot about python and I don't remember about vector/matrix operations. Is it possible to do the following (matlab code on the command line)

***

import numpy as np
x= np.arange(1,4)
y=np.arange(4,7)
print(x,y)
print(x+y)

[1 2 3] [4 5 6]
[5 7 9]
 
SanSanych Fomenko:

You're just not aware of R, plenty of forums, perfectly maintained, has a huge amount of useful literature for us both in the form of books and articles. I don't know anything like that in Python. The usual modern practice for formulating statistical thoughts is code in R, and quite rarely in Python.

If you take the statistics of the use of the two languages, the prevalence of them are about the same, but in Python there are a lot of users who write websites. If we evaluate usefulness of these two languages for trading, R has an undoubted advantage. Originally R (the paid prototype S of the 70's) was designed to solve exclusively statistical problems and was called so: the system of graphs and statistics.


One last thing.

What to pay attention to google. Take Microsoft. R is part of Microsoft's software today.

Uh-huh, we take it. MS recently released a deep-learning toolkit freely available. For some reason it's on the pluses )) As they write, the goal was to provide maximum speed for speech and image recognition.

https://github.com/Microsoft/CNTK

 
forexman77:

I see, so I have completely forgotten the language ))

123456789
New comment
Reason: