# Link in my Stancon 2017 submission is down

#1

Hi,

I’ve just realized that the link of the appendix of my stancon submission is down.

I see it’s here:

(Sorry if it’s not the appropriate place, none of the interfaces made sense to create an issue in github)

Bests,
Bruno

#2

I couldn’t tell from this what needed to be fixed.

Is there a web page somewhere with a broken link? If so, can you give me the URL of the page with a broken link and what the link is supposed to be?

If it is for the Stan web site, it’s hosted on stan-dev/stan-dev.github.io.

#3

Sorry if it was confusing.

The broken link is “http://mc-stan.org/events/stancon2017-notebooks/kfold.html”, the link should have kfold.html of my submission:
https://github.com/stan-dev/stancon_talks/blob/master/2017/Contributed-Talks/07_nicenboim/kfold.html

Thanks

#4

Here’s what I need:

• URL of page with broken link

I still can’t find the URL. I tried this:

~/stan-dev/stan-dev.github.io(master)$grep -R kfold.html . ./events/stancon2017-notebooks/stancon2017-nicenboim-vasishth-retrieval-models.html:<p>However, PSIS-LOO can be affected by highly influential observations; the estimated shape parameter <span class="math inline">$$\hat{k}$$</span> of the generalized Pareto distribution can be used to assess the reliability of the estimates, with <span class="math inline">$$\hat{k} &gt; 0.7$$</span> indicating an unreliable calculation of <span class="math inline">$$\hat{elpd}$$</span>. Even though the calculation of PSIS-LOO shows several <span class="math inline">$$\hat{k}$$</span> above <span class="math inline">$$0.7$$</span>, and k-fold cross validation is recommended instead, for ease of presentation we show the results using PSIS-LOO. We verified that K-fold cross-validation <span class="citation">(Vehtari and Ojanen 2012)</span> with <span class="math inline">$$K=10$$</span> yields virtually the same results, see <a href="kfold.html">Appendix</a> for details.</p> ~/stan-dev/stan-dev.github.io(master)$


Does that mean it’s a link inside another document?

@betanalpha or @syclik — I think you guys put this togehter—do you understand what Bruno wants us to do here?

#5

Hi,

The broken link is under stan-dev.github.io/events/stancon2017-notebooks

We verified that K-fold cross-validation <span class="citation">(Vehtari and Ojanen 2012)</span> with <span class="math inline">$$K=10$$</span> yields virtually the same results, see <a href="kfold.html">Appendix</a> for details.</p>

and should be (are hardlinks ok?)

We verified that K-fold cross-validation <span class="citation">(Vehtari and Ojanen 2012)</span> with <span class="math inline">$$K=10$$</span> yields virtually the same results, see <a href="https://cdn.rawgit.com/stan-dev/stancon_talks/03774fb8/2017/Contributed-Talks/07_nicenboim/kfold.html">Appendix</a> for details.</p>

edit. changed the url from https://github.com/stan-dev/stancon_talks/blob/master/2017/Contributed-Talks/07_nicenboim/kfold.html to https://cdn.rawgit.com/stan-dev/stancon_talks/03774fb8/2017/Contributed-Talks/07_nicenboim/kfold.html
I’m not sure what is the correct link. This new link is created in with rawgit

edit2. I created PR in github