Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #97308 > unrolled thread

Generating a vector from successive multiplications of another vector from an initial value

Back to article view | Back to comp.lang.python

Contents

  Generating a vector from successive multiplications of another vector from an initial value Paulo da Silva <p_s_d_a_s_i_l_v_a_ns@netcabo.pt> - 2015-10-01 21:45 +0100
    Re: Generating a vector from successive multiplications of another vector from an initial value Ian Kelly <ian.g.kelly@gmail.com> - 2015-10-01 15:10 -0600
    Re: Generating a vector from successive multiplications of another vector from an initial value duncan smith <buzzard@invalid.invalid> - 2015-10-02 00:44 +0100
    Re: Generating a vector from successive multiplications of another vector from an initial value Paulo da Silva <p_s_d_a_s_i_l_v_a_ns@netcabo.pt> - 2015-10-02 02:01 +0100

#97308 — Generating a vector from successive multiplications of another vector from an initial value

Hi all.

What is the fastest way to do the following:

I have an initial value V and a vector vec of (financial) indexes.
I want to generate a new vector nvec as

V, V*vec[0], V*vec[0]*vec[1], V*vec[0]*vec[1]*vec[2], ...

A numpy vectorized solution would be better.

Thanks

[toc] | [next] | [standalone]

#97309

On Thu, Oct 1, 2015 at 2:45 PM, Paulo da Silva
<p_s_d_a_s_i_l_v_a_ns@netcabo.pt> wrote:
> Hi all.
>
> What is the fastest way to do the following:
>
> I have an initial value V and a vector vec of (financial) indexes.
> I want to generate a new vector nvec as
>
> V, V*vec[0], V*vec[0]*vec[1], V*vec[0]*vec[1]*vec[2], ...
>
> A numpy vectorized solution would be better.

That looks hard to vectorize since each calculation depends on the
previous. You might be stuck with something like:

result = [V]
for x in vec:
    result.append(result[-1] * x)

[toc] | [prev] | [next] | [standalone]

#97320

On 01/10/15 21:45, Paulo da Silva wrote:
> Hi all.
> 
> What is the fastest way to do the following:
> 
> I have an initial value V and a vector vec of (financial) indexes.
> I want to generate a new vector nvec as
> 
> V, V*vec[0], V*vec[0]*vec[1], V*vec[0]*vec[1]*vec[2], ...
> 
> A numpy vectorized solution would be better.
> 
> Thanks
> 

Maybe,

http://docs.scipy.org/doc/numpy/reference/generated/numpy.cumprod.html

Duncan

[toc] | [prev] | [next] | [standalone]

#97323

Às 23:36 de 01-10-2015, Oscar Benjamin escreveu:
> 
> On Thu, 1 Oct 2015 21:51 Paulo da Silva <p_s_d_a_s_i_l_v_a_ns@netcabo.pt
> <mailto:p_s_d_a_s_i_l_v_a_ns@netcabo.pt>> wrote:
...

> 
> V * np.cumprod(vec)
> 

Thank you very much Oscar and Duncan.

I googled a lot for such a function. Unfortunately the word "cumulative"
didn't come to my mind!

Paulo

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.python

csiph-web