<div dir="ltr"><br><div class="gmail_extra"><br><br><div class="gmail_quote">On Thu, Sep 19, 2013 at 2:20 AM, Brian G. Peterson <span dir="ltr"><<a href="mailto:brian@braverock.com" target="_blank">brian@braverock.com</a>></span> wrote:<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="HOEnZb"><div class="h5">
<br></div></div>
Ross,<br>
<br>
Thanks for doing this, it's looking very good.<br>
<br>
I'm impressed by the comparison of the different random portfolio methods. It's pretty clear that at fev=0.05 the simplex method will concentrate more towards the individual assets while the sample method will get a more even distribution in the interior and near the edges. It's key to note that depending on your objective, you may very well not have an optimal solution along the vertexes. For example, risk contribution or risk budget objectives will likely place the optimal portfolio somewhere in the interior. The min-ERC portfolio lies along a line between the minimum variance portfolio and the EW portfolio.<br>
</blockquote><div><br></div><div style>Agreed, I think it is very interesting to look at the feasible space with different values for the fev biasing. Good point about the portfolios with risk budget objectives being on the interior. I'll add that to the vignette.</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
I wonder why you're using all the page space you are to separate the weights plot from the scatter plot. Maybe once to demonstrate that they are separate functions, but it seems that when you want to display both right next to each other, just calling plot() on the output of optimize.portfolio would be s more efficient use of space. e.g. Fig 4/5, 6/7, 8/9, 10/11, 12/13, 15/16 etc.<br>
</blockquote><div><br></div><div style>I did it once to show that they can be plotted separately, but then got a little carried away with copy/paste. I'll edit and just use the plot function.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
I also wonder if most of section 5.2 could be separated into a separate vignette specifically discussing the different methods of doing random portfolios, the algorithms themselves, and more completely describing the differences, but that can always happen after GSoC.</blockquote>
<div><br></div><div style>This would be interesting to expand on the material in this section including some benchmarking, looking at different constraints, impact of handling constraints by elimination, etc.</div><div style>
<br></div><div style>Thanks,</div><div style>Ross </div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="HOEnZb"><font color="#888888"><br>
<br>
<br>
<br>
<br>
-- <br>
Brian G. Peterson<br>
<a href="http://braverock.com/brian/" target="_blank">http://braverock.com/brian/</a><br>
Ph: <a href="tel:773-459-4973" value="+17734594973" target="_blank">773-459-4973</a><br>
IM: bgpbraverock<br>
______________________________<u></u>_________________<br>
GSoC-PortA mailing list<br>
<a href="mailto:GSoC-PortA@lists.r-forge.r-project.org" target="_blank">GSoC-PortA@lists.r-forge.r-<u></u>project.org</a><br>
<a href="http://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/gsoc-porta" target="_blank">http://lists.r-forge.r-<u></u>project.org/cgi-bin/mailman/<u></u>listinfo/gsoc-porta</a><br>
</font></span></blockquote></div><br></div></div>