"Identifying Unclear Questions in Community Question Answering Websites" by @jantrienes and Krisztian Balog accepted as full paper at the European Conference on Information Retrieval.

@s2205270 @hiemstra you can also try running docker-compose without the -d flag to see the startup logs

@assbach @Luca Über diesen Weg habe ich vor ein paar Monaten ein Sample von dem Mastodon Graphen erstellt der 250k Nodes beinhaltet. Ich kann den Datensatz gerne teilen wenn jemand interessiert ist.

@assbach @Luca man kann über die Web „API“ die followers jeder beliebigen Instanz abrufen: <instance>/users/<user>/followers.json

Ist aber durch das page Limit von 12 sehr zeitaufwändig.

The Mastodon API has a new endpoint for retrieving follow recommendations based on triadic closures.


Hi, we generate personalized follow-recommendations for you.

Sounds interesting? Then follow this account and you will receive a set of users that you might find interesting, very shortly.

@BeerendGerats How do you plan to deliver the recommendations? We noticed that when you send private messages containing a list of recommendations in form of mentions, all mentioned users receive that message which is of course not really useful.

@bouke This website looks useful! I wonder when instances get listed on it. I also found that you can get a list of "peers" from the API of an instance. For example: mastodon.social/api/v1/instanc gives a list of ~4600 mastodon instances. However, a large parts of them seems to be offline.

Our selection for the paper presentation:

Wang, Tianyi, et al. "Understanding Graph Sampling Algorithms for Social Network Analysis" (2011).


@hiemstra It seems that either way, @andrestorres and I would have to skip parts of or . We will ask Ansgar if we can move our presentations to the first slot (09:45), so that we will be in time for the original slot at 10:45.

Some discussion material for the progress meeting:

I crawled the topology of :mastodon: :utwente:. Red nodes are users from other instances. Edges indicate a follow relation. This is all the data we can get from our API. Will it suffice for a "whom to follow" recommender, or do we need to crawl other instances?

Interestingly, @hiemstra is the centre of our fediverse.

@hiemstra The original link still doesn't work. Only the external one.

Mastodon U. Twente

A social network for the University of Twente community. Anyone with an @utwente.nl or @*.utwente.nl email address can create an account here and participate in the global fediverse with millions of other users. This means that students, staff and alumni can create an account here. Content does not reflect the opinions or policies of the University of Twente.
We support \( \LaTeX \) formulas: Use \( and \) for inline LaTeX formulas, and \[ and \] for display mode.